Guitar Amplifier
Guitar Amplifier
Are you deaf (?)
Well actually yes! I had severe measles when I was 8 and went blind and deaf for 3 weeks -my sight did return as did most of my hearing. Then at the age of 15 I had to have tumours removed from my mastoids -that robbed me of some more of my hearing... But despite this I went on to buy and enjoy my LP records, CDs and SACDs. I went through a number of items until I arrived in 1984 at the “perfect” (for me) system to listen to music with... It consists of a pair of QUAD electrostatic speakers, the ESL 57, powered by a VTL Stereo 50 power amp (made in the UK!) this then has a Croft “Stereo Integrated Pre-Amplifier” (inside of which is a serial number (24) and a sticker with the date on it signed “OK -Glen”). The record player is SystemDek 3 by Dunlop -this is coupled with an Origin Live modified RB300 arm and their DC motor kit in which sits a Goldring Elite cartridge.
Over the years I have added and a CD player and a SACD player -and I still listen to the tapes made during my student days in the late 70’s on my Ferrograph... (Oh what do mean we weren’t supposed to plug our main XLR feeds into your pre-amp outputs from the band on the stage? Oh sorry that was just a mistake on our part -it won’t happen again. Honestly....)
They say that first dates are difficult -my to be wife commented on my system asking “What is that heap of Junk in the corner”? (Pause) “That is my music system”. I dusted off the LP of choice. I played “Wake of the Flood” and played it at “low” volume. The next date she came armed with a collection of “Club” and “Trance” CDs....
The rest as they say is history!!!
We now have a soon to be teenage son and for his Xmas present this year we bought him a (cheap) Fender Stratocaster and a (copy) of a “Tweed” amp. This has awakened things from my teenage days and now that several projects regarding the railway have had to stop -the roof of the “shed” has been removed and a new one is being built for it -thus all the “shed” is in boxes and tarps and will be for a few months yet until everything is finished. Mills & Lathes etc do not take kindly to being covered with brick dust and plaster....
So I am going to build something that “The Teenager” (me) designed when he was 15 as the “perfect” guitar amp. This may seem strange but the teenager eventually got to design liquid cooled super computers -thus a guitar amp might seem a slight diversion -but not so! In the late 60’s and early 70’s I was very poor. I had to experiment on electronics from hand me down and scrap equipment and carefully dole out my 3/6d pocket money on Essentials Only. At this time there was one thing that was “easy prey” that was a 405 line B&W TV set. The EHT system was a complete No-No to play with -but the audio side was a gold mine.The standard valves that could be found and used were the 6V6GT and the 6SN7. The former an audio output pentode valve the latter a double triode used for the scanning line system. The audio transformer was ripped out to power a small 8 ohm speaker (what do mean impedance matching?!?!?!?) The std setup was to use the 6SN7 as a phase splitter and the 6V6GT as the push pull driver valve. A small number of these were made for the local teenage musicians and I suppose they moved on to bigger and better amps...
I do have some favourite valves that have stuck with me from my teenage days. These are; the ECC81, the 6SN7, the 6V6GT, the EL-34 and deep respectful pause, The KT66.
So, what sort of amp am I going to build? The Pre-Amp will use ECC81s and the Power Amp will use a 6SN7 driving dual KT66 in a Parallel Single Ended triode mode. This should give me 10 Watts of Pure Class A1 and an awful lot of heat...
It has been a rather large “culture shock” to my system that the things that were oh so easy to find for “The Teenager” now are rarer than a set of orthodontic braces for a hen... But thanks to sheer bloody mindedness and the internet, I have been able track down quite a few of the components that I want -even if the manufacturers are now extinct(!)
I have dubbed my project “Cyber-Steam” as the amp as envisaged will be a hybrid of “modern” computer grade components (I have the left over bits from working on a Control Data Cyber II) and NOS stuff from the steam ages. The design would not have been out of place in the late 40’s and 50’s.
The first thing I had to track down were the capacitors. The type I wanted are called “Oil and Paper” and are about the size of lipsticks. I knew what I was looking for -and I purchased the entire stock of Farnell in one swoop -this gave me 4 of them.... After the end of some days of searching NOS suppliers I nearly have enough of them.
This is going to be a hybrid amp in more than one respect. The input to the amp has a 5 parallel JFET booster -the idea of this is that as noise is truly random the noise from each JFET will cancel -thus providing a quieter background. A similar system is employed in the Ortofon MCA-75 Moving Coil head amp that I use.
Although the circuit diagram above shows the 5 sections of the guitar input circuit as a whole, in fact it will be arranged as 5 circuits in a radial fashion. This boosted signal is then split three ways. The tone control system consists of a Peter Baxandall style Treble and Bass with a Presence and Shaper control.
How I work this -is so:
The boosted signal meets one side of the ECC81 and is then fed to to tone stack and is then returned to the OTHER side of the ECC81 thus any losses from the tone stack are re-amplified. This is then fed into an ECC81 both sides running in parallel as a mixer valve and the output signal then feeds both sides of a 6SN7 used as the driver and then the signal feeds the KT66s. (Thank you Mr Williamson).
The first thing you may ask is, why am I running double triode valves in parallel? If I run both sides of a double triode (the ECC81 and the 6SN7) I lower the noise floor, (useful), and I double the transconductance, (very useful), and I push the overload ceiling higher, (useful), and I lower the impedance of the source, (VERY VERY useful)...
The power supply for the amp is in two halves -one designed for the power amp and the other for the pre-amp side of things. The power amp side used Schottky Power diodes as will the driver and the pre-amp side. But, I do have to ask myself is -would a GZ34 be simpler? The heater side of the amp will be DC at 12.6Volts for the ECC81s and 6.3Volts for the 6SN7 and KT66s. The use of DC will reduce the hum and the use of 12.6Volts in the pre-amp valves will reduce the amount of current that is running around the chassis. This will reduce the current from 300mA to 150mA per valve. This is also easier to manage with thin wire star cabling -but at 1.3Amps each the KT66s and the 6SN7 will still need to be heavy duty wired...
The main advantage of a GZ34 as a rectifier is that it does not have supersonic ringing problems in the transformer core that a standard silicon power diode would have -thus the use of Schottky power diodes with capacitors across them in the power amp side of things. The “DC” provided by a valve rectifier is more “fluid” than that provided by silicon and as (I believe) the pre-amp is the most important “tonal” part of the system. I learned this “fact” when was “shopping around” for my music system in the early 80’s. Up until hearing the Croft “SIP” the pre-amp that I had used as my reference was the Pink Triangle “PIP”. After just 2 tracks it was obvious that the Croft “SIP” produced MUSIC whilst the Pink Triangle “PIP” produced HiFi.
It has also been said that the problems with a valve amplifier start and end with the transformers...
This is unfortunately absolutely and totally true. The mains power transformer has to have multiple windings -to supply all the various heaters, HT etc and this means lots of core. The output transformer that I require is a single ended one. These have a “gap” in the E&I laminations to prevent saturation. I also intend to fit the 6SN7 with a “plate choke”... As to the maths involved in this -there does not seem to a fixed solution(!) The basic “rule of thumb” is to fit a choke with more than 20 Henries and twice the current rating of the valve that it is going to feed. I intend to fit a generic resistor to the anode plate and then a variable 3 watt wire wound one before the choke to the anode plate. This will produce a voltage lower than the 250V that I require and I can “dial in” the required voltage. I then take apart the resistor and wire wound -then measure them to within a milli ohm of their lives and fit a standard one in their place....
The chassis that I plan to use is a commercial one from Hammond and is of course an all Imperial measurement part (17 x 10 x 4 inches). This means that all the rest of the things carefully placed on it are in metric -but have to have Imperial centred spacings... I bought some valve plates from E-Bay and the dimensions of the Chinese made part are all in metric -there are exactly 3 plates across the chassis and 6 plates along it -is somebody having a laugh at my expense?!?!?! I have some nice new Chinese made ceramic 9 pin sockets to fit the plates and some McMurdo (original) brown phenolic octal valve sockets for the rest. How do I know they are original??? Well the bag they came out of had “2/4d each” written on it.... My Father did business with His Father -it is that sort of parts store!!!
(I have a feeling that my son will do deals with his son).
The main problem I have using a steel chassis is the fact that steel is magnetic -thus I will need to elevate or insulate the transformers from the steel. I plan to use 15mm anti vibration motor mounts that will both raise the transformer beyond the inductive plane and help stop conducted noise from the windings. The cooling of this pet monster will be via fans -I plan to fit them UNDERNEATH the transformers -thus the transformers are force cooled by the the suction. The venting of the air will be via the valve mounting plates which have nice cute 8mm holes arranged around the socket fittings. This may seen a little bizarre but the reason for it is that in a “Pure Class A” amplifier -everything is running at full power all the time. Thus in a normal AB class amp the amp is only running in class A for a small amount of signal then changes to class B for the higher volumes. I am also running the KT66s in TRIODE mode. Yes this is a 30Watt power beam tetrode valve being crippled to produce at most 5W.... This is not like my beloved VTL, a HiFi power amp -but more a musical instrument. Why did I choose KT66 over EL34? I have found the KT66 to be very relaxed and the EL34 to be “aggressive”. The VTL is RATHER aggressive and I remember remarking that it had more attack than “being hit with a 6 foot length of 6 inch steel pipe”. It also plays Bach and Beethoven very well and can be incredibly delicate on female voice -but it is not the type of sound I want from my amp....
What is going to “feed” this amp is modified “Les Paul”. Yes -I play the blues!!!
“Mr Paul” has active neck and bridge pickups and the tone capacitors are oil and paper -but the metal in them is TIN -not Aluminium, and he is wired with silver coated copper cable. The control cavity and pickup slots are lined with copper foil and the four controls are 12 way “make before break” rotary switches using 1% metal film resistors. Yes -a PAD attenuator!!! Then, to add horror upon horror, the TEN control switches that will front the amp will also operate on the same principle. (As a piece of personal humour this is going to be an amp that you really CAN, “Turn up it all the way to Eleven!!!”)
Treble Slope. Treble Volume. Treble MIx.
Bass Slope. Bass Volume. Bass Mix.
Presence Shape. Presence Volume. Presence Mix
Master Volume.
Why a Presence control? Well I feel that the Presence control is more accurate, (musically speaking), than a mid control as found on other amps. The controls also reflect my personal taste in that they are SLOPE not cutoffs or boosts -yes I have had a QUAD 33 pre-amp and it leaves a long memory... The Presence has a shape control that alters the affected region from a plus 6dB parabola to a minus 6dB parabola in the band 1.5kHz to 4.5Khz.
I spent a couple of days at the spreadsheet working out all the resistor values for the ten PADs and I am proud to say that all the steps fit to within plus 0.43% to minus 0.29% of the correct values using two resistors. As this is below the 0.5% tolerance of the resistors I feel reasonably content with the PADs. Oh -this is also going to be a Left Handed Amp. Not because I am left handed, but because I am fed up with having to set the controls of other peoples amps with my right hand, (I like to strum while setting the controls), thus my right hand is busy while the left hand is trying operate right handed controls. Thus to me it makes sense to build an amp with controls that are left handed.
Some of you may have noticed the BC108c transistor at the base of the Cathode in each half of the ECC81. This is acting as a Constant Current Source. The modern method is to use a red LED but this does not appeal to me. The reason is the last thing I want to worry about is ; “Is that red light coming from the bottom of my valves just the LED or is something glowing red hot that I cannot see”? The use of a Constant Current Source in a guitar amp is uncommon to say the least! The reason I use this approach on the pre-amp is to lock everything into a fixed mode and as a follow on -it is quieter too. Depending on how I feel once I have heard the resulting assembly, there may be a small polypropylene cap across the transistor to pass any errant AC.
There will be two systems of “ground” in the amp. The first one I am calling the Power Ground. All the power connections will (as far as possible) star earth to the negative pole of the first capacitor after the bridge rectifier in the main power amp PSU. This will then have a small wattage 10ohm resistor from it with a couple of polypropylene capacitors across it and connect to the main chassis earth point. This small filter section should help remove any RF from the PSU side. The other ground is the Audio Ground. I would love to be able to star earth this -but I don’t think that it is physically possible. The underside of the amp would resemble a explosion in a wire store.... So, the Audio Ground is going to have to be a “Bussed Star” arrangement. The end of the Bus Wire is going to be the pin 1 of the male XLR socket, this then earths to the chassis earth point by a similar filter as used in the Power Ground. YES -I am using XLRs. I far prefer them to phonos or 1/4 inch jack plugs. The output from the amp will also use an XLR female socket as will the pre-amp. This reflects my personal viewpoint and should not be taken as “gospel”, but I think the XLR is an almost perfect audio connector in that it has built in shielding and always makes “earth” before “signal” connections. As a bonus if you tread on them -you don’t hurt anything. The other one I like is the 3 pin DIN -but modern day DINs are not what they once where....
The next picture shows the first rough idea of what it will look like. Some of the parts, like the GZ34, I am still not sure about(?)
The next shot shows the valve plate and the stuff that will be assembled around it. The Paper and Oil capacitors are obvious as are the polypropylene capacitors, the cermet multi-turn trimmer potentiometer is rather small and I did check with the suppliers that I had got the right ones! This will be repeated 3 times for each part of the tone assembly -the mixer stage has a few more capacitors...
The next shot shows the assembled glass cast... They are, (from left to right), a French made GZ34, a Slovakian made ECC81, a Russian made 6SN7, and Canadian made KT66 from that GE factory in Quebec. How do I know? The box had details in French then English...
Well a busy week and a lazy Sunday afternoon to work on the amp in the kitchen. The chassis and base plate are made by Hammond, and I got them from BlueBell Audio in Scotland. “Phillip” the mgr had a look for me at his stock -thought the two that he had in stock were not “of merchandisable quality” and ordered me a new chassis and base on the phone from Hammond in Canada(!) This arrived on Friday Morning and it looked simply beautiful!!! I admit to feeling rather awful at having to drill and cut holes in its grey perfection -but this is what has to be done.... I had marked out my plan on A3 Paper and I stuck this to the top of the chassis with tape. I then placed the valve plates on the paper and then looked at it.
Something was not quite “right”.
I had all the plates edge to edge and that is what seemed to be wrong. I moved the plates until there was a 5mm border of grey steel around them and -that seemed to be “right”. Moving the first of the KT66s to the centre line of the chassis and then positioning the second KT66 equidistant from the edge as the 6SN7 also improved the feeling of “rightness”. SO, that is how I have drilled it...
The pilot holes I hand drilled with a wheel brace using a 1.5mm TiN drill. TiN drills are expensive but considering how much tougher they are -I think they are worth it. The outer holes I used a battery drill and a 5mm Slocombe double ended drill. I find these to be better for working with sheet materials as the small Bell Slocombes rarely slip on the material as the the leader punches through and this holds the main part of the Slocombe in position.
The next stage will be to dig out the tank cutters and then cut out the vent holes for the valve plates. Then to get the threaded insert adaptor for my rivet gun, (I am sure that it is somewhere very safe -as I have yet to find it!!!)
It is now Tuesday morning and I have yet to cut the holes for the valve plates -the reason is I cannot get any of my hole saws to cut through the steel plate... This means that I have had to scour e-Bay and purchase a suitable tank cutter. The Bank Account reveals that other parts that I ordered over the week end are on their way. These are the blue anodised valve shields, fuse holders and switches that I need for the amp. There are a few “fripperies” amongst the order too. I have bright red swirly circuit board material to fix my blue turret tags onto and nice bright blue pilot light. (To quote my wife; ” Do I detect a BLUE thing going on here?”) I have also some 11 Amp solid core cable that will be used to produce the valve heater circuits from. I can see quite a lot of wire twisting going on in the kitchen soon! The yellow cable will carry the 6.3 Volt heater current and the violet the 12.6 Volt heater current
The first thing to install in any wiring circuit in a valve amp -are the heater circuits. These are “traditionally” shoved into the corners of the chassis so the amount of electrical hum that they generate is minimised. I am going to use DC heaters and I am going to use earthed braided shielding tube to contain them in -thus the amount of noise flying out of them should be minimal, (I hope!) I also plan to wire the internals of the amp in “balanced” mode. This means using a shielded cable with TWO conductors in it. I earth the shield to the ground point and the signal wire feeds to the device and the drain wire feeds to the local audio star ground. The only really suitable cable for this is microphone cable.The problem with this is that I am not exactly sure how much heat the cable will take...
Well in the evening I did a rough dry assembly of the top of my amp to see if I had the correct “feel” to it.
I liked what I saw. The red circuit board material is there as a intermediate measure -it is going to be replaced with brown paxolin sheet cut into a T shape. The red circuit board material does look nice -but being loose weave fibre glass it does leave itchy bits all over your hand when you work with it... The next shot shows some of the switches and other frippery that I will encrust the exterior of the amp with. OK -so the switches may be “slightly overkill” at 15Amp rating, but they have a very nice solid “clunnnngggg” as they flip over. I have a feeling they will last the course of the amps life....
There are two external fuse fittings. One will fuse the power amplifier and the other will fuse the pre-amp. There will be a fuse (as per UK regulations) in the mains plug. I intend to use Belkin mains cable with a shielding foil and drain wire. These will be earthed at the mains plug end of the cable.
Well at the end of another Sunday afternoon in the kitchen I have retired to the living room with my computer a block of butter and large stack of pyclets.... Well can now say that all of the holes have been pilot drilled -the nice greyness of the chassis now looking as if it has had a blast from a shotgun. The next shot shows the cut sheet of paxolin that will cover most of the top of the chassis. The holes were transferred from the chassis to the paxolin sheet by taping the sheet to the top and cutting the two initial holes gently with a 5mm end mill. Once they were in place I bolted up the top plate and simply blasted the rest through! On the grey side of the chassis you can see the pilot holes for the two main power capacitors and the socket for the GZ34 -yes I have finally made my mind up.
Flipping the chassis over onto its front shows the positions of the ten control switches, (top row), and the bottom row is as follows: Pilot lamp, “Dirty” XLR , “Dirty” Gain potentiometer, “Clean” XLR, Mute on/off, Negative Feedback on/off.
The rear view shows the following on the top row: Fuse P-Amp, Fuse Pre-Amp, Main Power and Air Conditioning switch. Yes I know this sounds a little strange but the amp is going to need to cool down once it has been switched off. This will power the cooling fans. The bottom row is the output XLR (although I am looking at a more modern plug -maybe a “Speakeron”), the Earth terminal and the main socket.
The next shot shows the test fitting of the partially completed circuit board. There are 27 rows of turret tags.... These are screw in type and require a 6-32 thread cut into the circuit board material. This was dead easy with a drill/tap fitted to my trusty screwdriver. Each one of them has been CA’ed solid. The weight of the 3.2mm thick board plus the tags was errrmm -quite a lot!!! To remove problems with board sagging I CA’ed three strips of pine to the back of it, thus the circuit board resembles a section of box girder bridge... It is supported at two points by M5 standoffs. These “back up” to the standoffs on the top of the amp
The next shot shows the board with some of the capacitors rough fitted. I tried weighing the assembly on my wife’s “diet” scales. The total weight of the board and parts (and there are MORE TO COME!) was a little over 425 Grammes. I am for future posterity and ease of alteration -mount all the capacitors with their labels UPPERMOST. This saves on the eyesight and memory(!)
The bottom plate of the chassis has been re-enforced with three lengths U shaped 15mm section Aluminium extrusion. This will help the base plate take the weight of the chokes and also provide a source of heat sinking for components that are TO220 devices. These I can simply line up over the Aluminium section and blow a 3mm holes through it. Hey Presto -a large air cooled heat sink fitted! The next thing to buy on the shopping list is some varnish. This may be surprising -but it seals the circuit board and (more importantly) seal in all the flying itchy pieces of glass fibre that are plaguing me at the moment!!!
Well it is now Tuesday morning after a very frustrating two days in the kitchen....
The plan was that Monday I would use the tank cutter (brand new) to cut the holes in the chassis with. Everything went well until I actually started to try to cut with it. I loaded the rank cutter into my pillar drill and mixed the belts until I had 350RPM. I hit the green button and (Rumble) everything went round well. I then tried to cut through the metal. The nice sharp steel cutter hit the paint and then -nothing... Oh -a nice shiny ring of polished steel appeared -but other than that -nothing... I applied a bit more “wellie” to the pillar drill and the ring of polished steel got wider -but no deeper. I hit the stop button and examined the cuter tip -it had worn away!!! A phone call to the supplier reveals the problem -”It is Carbon Steel. You are not supposed to use it on metal”.
I delicately replaced the telephone and then went out side and screamed!!!
I then went into my shed and rummaged around. The net result was some old lathe cutting tips in T/C and my Mapp torch and some silver solder. A few minutes later I had soldered a T/C tip to the tank cutter. This then lasted long enough to cut the fan hole (7.5cm) and the first of the valve holes (5.3cm) -then the screw holding the bar gave way and the cutting arm flew out and smashed the T/C tip...
Deep Breath(!)
I packed everything away and tidied up. Tomorrow (or Saturday) I am going to Machine Mart and purchase myself a set of diamond tipped cutters!!!
Well as everyone in Europe will tell you -we are in the grip of a deep freeze. Which does of course mean that no-one is going anywhere -except closer to the fire... While I wait for the ice to thaw I have been figuring out how to solve the one problem that has bugged me -how to safely power down the somewhat lethal power supply that is now fully charged. The idea is both simple and cunning. I use an industrial 2.8kVA 3-phase isolation relay. I power the relay coil from the 230V mains switch for the Pre-Amp. When the Pre-Amp is switched OFF at the panel switch the contacts close and the power stored in the caps is shunted to earth via a trio of 3.3K resisters -each rated at 25Watts. The fans are non-switchable, (i.e. they run when the amp is switched ON at the mains), thus the 25Watt resisters are cooled whilst they discharge the three capacitor banks. There is also going to have to be a “buss bar” from the three resistors to the main earth point. This does mean some clearly defined methods of powering up and powering down the system... In my domestic music system, (which is currently playing Marillion “B Sides Themselves”), I power up the turntable then the Pre and then the Power amplifiers. The shut down will have to be in the reverse order -otherwise the 3-Phase relay will close on a fully working PSU!!!
(I don’t think that the fuses would like that...)
Now that I have most of the mechanical sides, (bar the holes), sorted out I am turning my attention to the look of the amp as a whole. I had started out with the idea that it should be a “steam punk” amplifier, but to be honest -I don’t really see it as that anymore... It seems to want to be a piece of post-WW2 equipment instead. The colour of the paint that I have chosen for it is “pearl black” this is like a metallic hammered paint that would not look out of place on an RAF radio of the WW2 period. Added to this I have got two “modern antique” moving coil meters and 7 sided knobs and PADs it looks like hi tech equipment of the period. I will keep the brass ”butterfly” hinges that the equipment would have had alongside the “Bakelite”. (OK -so it is actually Paxolin but once it is varnished behind the cage work -will anyone be able to tell???) It will also have brown leather strap handles at the ends of the cage. This will give it something to be lifted by as well as being “period”. I did try and persuade a laser cutting establishment to produce some skeleton dials of the right size -but they said that it would be too delicate for them (shame!)
The relay has been fitted and it seems to me to resemble “Robbie the Robot” out of “Forbidden Planet” with the front relay fingers -but if it ever says “Attention!” I will be the first to scramble the switch to the KRELL laboratory!!! The irritating thing is after having studied the pinouts I find that I have mounted it 180 degrees out of whack -thus I am going to have to mount it in the reverse of the pictures below! This gives me the feeds to the rear and the outputs to the front -rather than the way they are now.
After some soul searching and thinking -I have decided to mount the relays (electrically ) “above” the dropper resisters as this will keep that part of the circuitry “dead” until the Pre-amp DPDT switch is turned off. Being a TPDT relay I use the N/O connection to shunt the PSU capacitors to earth, thus once power is applied to the amp the relay clicks over to N/C and the capacitors begin to charge up. The seeming crazy logic behind this -is as follows:
The two DPDT switches at the rear of the amp are linked so that once the mains connector is plugged in the cooling system whirrs into life and the relay shorts out the PSU caps on the three HT lines. It is the action of BREAKS the connection rather than MAKES that produces the destructive arc. So, as the power DPDT switch to the pre-amp is thrown the relay BREAKS the connection from “dead” lines -thus no arc(!) The power amp DPDT switch feeds from the pre-amp DPDT switch -this ensuring that the Pre-amp is powered on first...
Thus whilst the amp is plugged into the mains, but not active, the three HT lines are HELD grounded.
The mounting bolts, (Brass), for the TO220 diodes and voltage regulators have also been CA’ed to the base plate. The ones for the regulators are aligned with the Aluminium re-enforcing channel -every bit of fan cooled conducting metal helps(!)
The area of the base plate is starting to become pretty crowded -even though it looks empty at the moment. There will have to be SIX chokes and an AC transformer on it along side the power circuitry required for the valves HT and DC Heaters, (one running at 12.6V and the other running at 6.3V). There will be one 2A 12V regulator, (plus one silicon diode under the earth connection), to supply the heaters for the Pre-Amp and three 2A 5V regulators, (and two silicon diodes for each VR), for the Power Amp. The AC feed for the 12.6V heaters will be from combining the std 6.3V feed for the heaters with the 5V heater feed for the rectifier valve.
I have been highly alarmed at the lack of books that I have access to for designing the circuitry for the amp and pre-amp. I have to shamefacedly admit that the manuals that I am using are the M-O Valve Company Application report No.20 and my “A” level Physics books (I really must get around to returning that to the school one of these years...) Am I using a spreadsheet to work out the parameters with -er no... I have my much maligned and scratched “Toison D’Or” slide rule. Other kids had “British Thornton” Slide Rules -the “Toison D’Or” range came with a set of interchangeable slides for chemistry, physics etc! I admit it did seem rather like cheating to be able to slide a set of pre-marked power functions and chemical standards. But it still didn’t stop me using it in my exams!!!! I sat down with a good mug of tea and did some calcs with the rule. My eyes not being what they were -even with reading glasses(!) I summoned up the scientific calculator on my Mac and plonked the keys after 10 minutes an scribbling on paper I got an impedance of 7.9K for my pre-amp stages. Looking at the figure on the curser bar, it was a slight hair under 8 and it had taken me 20 seconds....
After a few very strenuous evenings work we have the final calculations for the pre-amp side of the amp finally decided. That may read as a puzzling sentence -but bear with me it will be explained. “Board One” has 27 turrets per side i.e. 9 per valve. These are spaced at 1.5cm intervals along the length. This board deals with the tonal amplification side -the treble, bass and presence controls. Thus it is a mixture of capacitors and resistors. Resistors are roughly 0.5cm square lengths of porcelain. Capacitors are any size or shape that they feel... So I have to calculate the type of capacitor I need for the correct curve for the cathode bypass capacitor. I have after several cups of tea -decided to gamble. Ideally the cathode capacitor should be 100uF to completely bypass any and all AC from the valve. But in practice -do I need this? Looking at the home made graph plots and twanging “Mr Paul” gives me a zone between 150Hz and 15KHz. Working on the known premise that the smaller caps favour the higher frequencies I sat down and (very painstakingly) worked what the “turnover” would be at various capacitor values. It would seem that 2.2uF is as large as I require(?) This puts it firmly into the polycap range of devices and I was dreading the problem of fitting large caps into the line of turrets.
The largest device on the boards will be the 47uF electrolytic “Local Power Cap”. It will not have much of a ripple to it -its main function is to hammer out that last fragment of noise on the HT line. I have also decided to star earth the pre-amp side using multiple stars and then running 0Volt lines to the Signal Earth.
That cast of “Board One” is as follows:
1: 22k 3W w/wound anode V1a
2: 0.01uF paper and oil capacitor (AmpOhm)
3: 1k 3W w/wound cathode V1a
4: 2.2uF polyester capacitor cathode by-pass V1a (Sprague)
5: 47uF Local Power Cap (Mundorf) -this rank is local HT feed and star ground points
6: 2.2uF polyester capacitor cathode by-pass V1b
7: 1k 3W w/wound cathode V1b
8: 0.01uF paper and oil capacitor (AmpOhm)
9: 22k 3W w/wound anode V1b
And this is repeated three times!
I varnished the boards and the lovely red swirl has come to the fore. It has also (nicely) put paid to the terrible itching that the fine glass fibre strands left as they stuck in my flesh...
The heart and lungs of the power amp have finally been hammered out as well. The output transformer will be a Hammond 1628SEA, not cheap at £90+vat &shipping -but at 5Kg no lightweight either!!! This is going to come from BlueBell. The power transformers are going to come from Danbury and the six chokes from Glass House. I have “The Vacuum Tube Logic Book” (2nd Ed.) and this black tome is rather good reading -if slightly biased in its outlook! I was going to produce a stabilised HT line but on reading the book, and thinking about the problems involved -I too had to agree that a simple Pi circuit is the best solution for the job. The old rule of thumb was that the number of Henries in the choke multiplied by the capacitance in MicroFarads should be greater than 220. Well looking at the numbers I think that I have it comfortably beaten -on the pre-amp side at least!
(100+100+47+47+47+47+47) MicroFarads x 25 Henries = 10,825
The supply on the power amp side is not quite so “rich” but is still a comparable figure.
(100+100+100+100+30) MicroFarads x 25 Henries = 10,750
The 30uF cap in the supply chain to the power amp is an “odd” one... It is a “TubeCap” by Mundorf. When I read about them I knew that I had to have one of them. This is a 30uF polycap!!! I don’t know whether this will improve the sound quality -but I want one!!!
I have also sat down and worked out how much “acoustic energy” the power supply for each part of the amp will have. The figures were initially depressing -but on some reflection possibly not as bad as they seem.
The standard power formula is:
Joules = Capacitance x Volts x (Volts / 2)
This gives me a Pre-amp PSU rating of: 435uF x 300 x (300 / 2) = 19.57Joules
With a Power Amp PSU rating of: 430uF x 450 x (450 / 2) = 43.53Joules
Consulting the “Vacuum Tube Logic” book shows that their largest amps have PSU ratings of 287Joules, but since these are vastly more powerful devices I am now not too displeased as the rating of PSU Joules per Watt output on my PSU is higher than theirs!!!
Over the past few days the parts have been drifting in from the suppliers, Maplin, Squires and Modulus etc. Saturday Morning was spent bashing metal -i.e. straightening out the leads to the components. These were then duly wrapped around the bases of the turrets and the result does look somewhat artistic, this was not intentional -it just happened....
The ugly space at the RHS of the lower of the two boards will be filled with two 22uF BiPolar electrolytics. These will act as the cathode capacitors and will be bridged with the 2.2uf polycaps. There is some debate as to effect that bridging BiPolars with polycaps have on the sound. My “take” on it is -that I don’t like it then it takes two clips with the cutters to remove the problem... The final parts of the power amp will be hard wired to the Octal sockets as this will ensure that they are in a cooling draught all the time -via 3cm fans blowing across them!
