Since I was building a 5F6-A Bassman, the amp that was copied to make the first Marshall, I thought it would be informative and fun to build that original copy to see what I could learn.
Once again the originals are crazy money, so we rolled our own. I was lucky enough to locate an original RS Spares Deluxe output transformer to use so at least that bit could be authentic. I am partial to output transformers, and this particular one is very interesting because it was originally a 30W general purpose HiFi model. It has a very high primary inductance compared to most guitar amp units, grain oriented core laminations stacked in 10's, and it has a variety of different primary and secondary connection options. It also famously appears in the "Bluesbreaker" amp. Oh, and it's painted in grey hammertone. #coolman !
I was making reproductions of these transformers at the time, and had been invited to a group of folks who were building these amps.. I got a prototype chassis from one person, a stuffed turret board from another, etc. It wasn't a kit, it was somewhat from scratch especially because of the unique output transformer mounting that requires a large cutout. Building something like this can make you appreciate all of the details you don't have to deal with when you build a kit.
Luckily the chassis was Aluminum which made the metal work much easier.
A Dremel with a cutoff wheel is a very effective approach for things like transformer mounting openings.
I had to enlarge the PT opening, this is prior to filing the corners.
PT is mounted, now it was time for the OT
Cutting the opening for the RS Spares
All the iron mounted
This was my big mistake. When the amp was done it was marginally unstable. The reason was the output transformer orientation. I didn't have a picture of the original, and at the time I thought that the primary leads were a bigger issue for instability, so I placed them closer to the output tubes. (the 9 connections side is the primary, the 6 are the secondary) And that was not correct; the secondary leads are the bigger issue because the current flow is so much higher. So later on after the amp was done I would have to go back in and flip this around. Thank you to Mitch Colby for catching this! (pic is how *not* to do it)
I always lead off with the power supply including heaters
Ray was kind enough to sign the board.
Test fitting to figure out the board support locations
Power supply is nearly done here
And our first power supply test.
Starting to wire up the board, once again flying leads were in place before the board goes in. This makes it easier to assemble and easier to do QC ahead of time.
The board, all in. Strictly speaking I should not have the ground buss on the pots, it's a ground loop, but the original had it and I just rolled with it. The amp is not noisy.
For comparison, here is an original JTM45. (Photo: Mitch Colby, I think!) Astute amp dogs will notice the OT in this one also has the secondary close to the control panel! Hmmm. It is wired for 16 ohm out on the secondary; I wish we could see the primary!
And here is my own 1970 Marshall 1986, metal front but the same layout and parts as a "Plexi". This amp needed a bunch of work but sounds amazing. It had a messy post PI MV with lots of goop among its issues. But the point is that the layouts are all basically the same; 50W, 100W, older, newer... There are capacitor changes in there for brightness, and a few smaller gain changes but nothing big. I'll do a page on it soon. It's pretty neat, and still rocks the very first OEI production OT!
It came out pretty well I think.
Output transformer notes - The RS Spares Deluxe
The output transformer in guitar amps is a particular passion of mine, which led of course to starting OEI and getting them designed and built. Among the things that make them fascinating is the apparent lack of a specific set of criteria that makes one good or better than others. Some are cheaply made, seem poorly designed - and sound great. Some, like the RS Spares, are pretty carefully designed and spec well. It used an interesting core assembly as well, 29ga grain oriented laminations, grouped by 10s. This is a technique to make the transformer handle imbalanced DC on the core better. Typical transformer cores are interleaved (some call it lacing) 1:1 (min) and of course SE OTs are butt stacked (max) and then in between, you have lots of other options. 2x2, 3x3, etc. It is taking advantage of the slight curve left from being stamped; they do not stack perfectly flat. You lose some primary inductance doing this, but if your bigger concern is core saturation effects, it may be well worth it. I am not sure if there assembly cost variances to it.
There were other guitar amp manufacturers who used HiFi outputs as well; Ampeg used Dyna very early on, as did Sun. Ken Fischer's Trainwreck amps used a variety of outputs, (including a Stancor with very interesting core stacking..). And here, with the new-born Marshall company, they likely did not have the resources or expertise to design their own and have a bunch made. It was probably much easier to buy some off the shelf part at least in the short term. Taking 2 steps back, we see that Ken Bran built in both a lot of negative feedback, and used a very nice output transformer. Combined with increasing the first stage gain - which is generally a professional sound approach to minimize noise - was he really shooting for a bigger, more fidelic clean sound - and then Pete Townsend came along and did the unthinkable? Who knows. It is fun to wonder. It is a curious mix of design decisions when I look at it. And the results, intended or not, are certainly legend.
Here is an RS brochure. The "Deluxe" wound up in the JTM45 and possibly (anyone really know?) some very early AC30s (I have never seen one, or a photo of one but have heard rumors), & I think the "EL84" is what wound up in the Vox AC15.
The 'Deluxe' was pretty clearly targeted at Williamson amp use, with its 6k6 primary and UL taps with a 30W power handling. Which is one of two real issues with it. Issue #1 is arcing between the pins because they are so close, (see photo) and issue #2 is the low primary current rating and relatively low power output rating. It is rated at 30W, and we should probably assume that is "clean". Now the wattage rating in an output transformer is really one part frequency , and one part wire gauge. [FIXED - not 1/2 power per octave]
You can see for that last 10 Hz, just 1/2 an octave more, you lose 50% of your rated output power. The core can't handle it. Now 50% may seem like a lot but it's the 3dB down point so you get to say your output transformer is good for 30W from 20-20,000 Hz +/- 3dB if you want to. Dyna was more transpaperent here.
I am citing them because they quoted it publicly and it makes it easy to digest. Tangentially, this is why aircraft power runs at 400Hz. Back when, they needed to save weight, and this saved a lot of weight from transformer cores in the equipment.
The 30W rating on the Deluxe is spec'd at 30Hz (stated above) The lowest fundamental on a guitar is 82.5Hz (low E) so the core is arguably fine past 30W, so 45W should not really be an issue here. What is an issue though, is the 125mA max DC current. A pair of KT66's are ok, but a pair of EL34's will push close to that. You primary wire is skinny, and buried under layers of insulation and other wire. It cannot get rid of heat very well, and so is typically de-rated to operate in these devices.
KT66 operating specs:
So in the JTM45 where we do not use the UL connections, we're just pulling plate current through the primary which is 62mA max here with the stock KT66s, which seems fine, clean. But how are things under clipping? Or tube failure? And what happens when someone swaps out the KT66 for some other tube? Not really a lot of margin for error.
EL34's are pulling around 106mA at just 400V, by 450V and in clipping - who knows. If a tube shorts, what kind of margin do you have? Do you feel lucky, punk? Just something to keep in mind. It's an expensive experiment.
EL34 specs are below:
So at max signal, our EL34's are pulling a total of 127mA, although 21 is on the screen. But 106mA is is a bit hairy considering our max, and considering most guitarists will be beating the heck out of the amp. Simcha was strongly of the opinion that EL34's would almost assure a shorter life for the OT. YMMV, but keep it in mind.
Nice specs: 67H primary inductance is clearly aimed at HiFi use. Leakage inductance of 21mH is also quite good. They are claiming 30Hz to 35KHz -3dB or better. Measuring this primary inductance out of operation is pretty tough; it depends on excitation current in circuit.
Measuring the RS out of circuit on my BK 847 handheld showed a primary L of 15.5H as an example. You're generally better off putting the device in circuit and measuring your frequency response at full undistorted output. You can then back into the primary L. The amps 3dB down point can be calculated with f = Z/(2*Pi*L), where Z is the reflected impedance on the primary, and you are solving for L. f is the frequency where you are 3dB down (half power point) from your maximum undistorted output at say 1kHz.
Now that part that drives guitar players mad but pleases transformer designers. The secondary windings are 3 separate windings that are combined in series or parallel to create the turns ratios required for 3, 7 or 15 ohm out. Switching it is pretty tricky.
To mis-match it (as I recommend), you wire it for 16 ohms and then run an 8 ohm cabinet. With the 8K primary (4-5-6) that will give you 4K and what I find (ymmv!) is the best overall tone.
Why does it please transformer designers? For one thing, you always want to use the whole secondary. There are inductive effects. Your leakage inductance typically gets worse using less than the whole winding; you are reducing the area it has to be near the primary. Think about the McIntosh OT with it's bifiliar winding which is kind of the best case.
Using the typical modern day setup where you have 1 winding and take taps off of it is not a great engineering practice. Is it audible? Does it matter? That probably depends. Fortunately perhaps, guitars have a very narrow passband compared to HiFi, and that almost certainly helps out.
What it looks like in practice. This is where I started my build out; 6k6 primary (1-5-9), 16 ohms out and mounted with the secondary towards the control panel. I later flipped the OT around, kept the 16 ohm out but wired the primary for 8K and then mismatched it. This cured the instability and improved the overall tone a good bit.
The 3 secondary windings are (start to finish) E to D, C to B and A to F.
I did a bunch of research and perhaps more importantly, tried a bunch of things out. Where I landed was I think the sweet spot is to use the 8K primary setting and then mis-match it so that the primary sees 4k. This is what I believe went on in the "Bluesbreaker" combo. It certainly sounded the best here. The heads seem to have been wired at 6k6? See what you like best! If you have conflicting info, I'd love to hear it! I am just guessing and using my ears. If it needs saying, ICBW!
References:
The old OEI website (available on the wayback machine)
Marstran, who had much cleaner versions of the RS docs than I did (thanks!!) http://forum.marstran.com/viewtopic.php?f=20&t=128&sid=fffd94bde33eed825b1cafd8907f2237#p750
Graydon Stuckey for driving the JTM45 amp project, including manufacturing the chassis, PT & choke i used. Thanks!
Ray Domzalski who built a heck of a board! Thanks!
Comments