My take on Eric Johnson's Steel String Singer #005

[martin manning]

Putting a new preamp on an existing power supply/power section seems to be a common theme, and why not? Another example is the lifting of the Music Man power supply and power amp, but in that case the power stage was converted from the original SS hybrid to all-tube.

[rootz]

There is something going on around the output of v1b. I noticed a 1meg resistor right of the plate of it leading to the filter network caps. You can see that in all of the pictures from AN amps shared here and on his Insta. Looks like a substitute for a maxed out trimmer on e.g. the 002. Seems logical as this amp has a rather small cathode cap on v1b.
What also catches my eye, is that the cap leading to the dwell pot seems to be in series with another. First a bigger (mica?) cap and then a coiled wire to a smaller ceramic cap (ceramite?) Both are also after the 1 meg resistor I mentioned above. I'd predict that the 1meg res from the v1b plate makes the filter section much more effective, more mid scoop too.
Another thing I can't spot in all pictures of AN 005 based amps I have are the 220k mixer resistors. Where are those?
Other differences:
- some amps do not have bypass caps on the mixer and reverb return tube;
- values on the PI differ a lot by the looks of it, tail resistors not typical 24k on some, bias resistor not 820r but 470r on others;
- there is a 220k resistor next to the PI components on the board on pictures I have. Looks like the global nfb resistor. That would be a lot less nfb that a 47k!
- next is the cf driving the outputs. 100k cathode resistors, but grid leaks and stoppers seem to be smaller. 150k stopper, 330k leak. Also coupling caps seem smaller than 47nF, or the coupler to the volume pot might be bigger than 22n.
Can't figure the PSU out good enough on pictures, but probably many differences there too.

Yes, I noticed that this isn't supposed to be a clone of a Bludo or AN. The network and values right after v1b and the filters seems critical to get right IMHO. part of where the magic happens I think.

[martin manning]

Rootz, if you put a 1M in front of the filter nw you create a low pass and kill the high end. Not a good idea, IMO. That 1M might be feeding the reverb drive pot to adjust the available drive signal level.

[beasleybodyshop]

There is something going on around the output of v1b. I noticed a 1meg resistor right of the plate of it leading to the filter network caps. You can see that in all of the pictures from AN amps shared here and on his Insta. Looks like a substitute for a maxed out trimmer on e.g. the 002. Seems logical as this amp has a rather small cathode cap on v1b.
What also catches my eye, is that the cap leading to the dwell pot seems to be in series with another. First a bigger (mica?) cap and then a coiled wire to a smaller ceramic cap (ceramite?) Both are also after the 1 meg resistor I mentioned above. I'd predict that the 1meg res from the v1b plate makes the filter section much more effective, more mid scoop too.
Another thing I can't spot in all pictures of AN 005 based amps I have are the 220k mixer resistors. Where are those?
Other differences:
- some amps do not have bypass caps on the mixer and reverb return tube;
- values on the PI differ a lot by the looks of it, tail resistors not typical 24k on some, bias resistor not 820r but 470r on others;
- there is a 220k resistor next to the PI components on the board on pictures I have. Looks like the global nfb resistor. That would be a lot less nfb that a 47k!
- next is the cf driving the outputs. 100k cathode resistors, but grid leaks and stoppers seem to be smaller. 150k stopper, 330k leak. Also coupling caps seem smaller than 47nF, or the coupler to the volume pot might be bigger than 22n.
Can't figure the PSU out good enough on pictures, but probably many differences there too.

Yes, I noticed that this isn't supposed to be a clone of a Bludo or AN. The network and values right after v1b and the filters seems critical to get right IMHO. part of where the magic happens I think.

- Martin's guess on that 1M resistor for the original design after the plate was to keep the Dwell control from loading down the filter circuit, or to prevent the filter circuit from loading down the plate driven stage output. I have no intelligent comment here so i'll leave it at that

- Yeah I did find the pair of caps in series to be a little odd. IIRC it was a 4700pF and a 270pF in series? I think that gives you like 250pF or so?

- I am pretty certain that 100K/220K resistor near the PI is indeed the GNFB resistor. It is definitely odd! But could explain why EJ liked it for a "dirty rythmn" amp. Maybe channeling some JTM45 style gritty vibes?

- Lions share of the output stage were taken from changes and improvements that people on here made, based on MHartmans take with Martin's wrinkle for improved cathode follower supply.

It's my opinion that the lack of a choke, and very high NFB resistor/less GNFB play the biggest differentiating factor from all the other well known iterations. Looking forward to piecing this one together and seeing how it shakes out!

[rootz]

Rootz, if you put a 1M in front of the filter nw you create a low pass and kill the high end. Not a good idea, IMO. That 1M might be feeding the reverb drive pot to adjust the available drive signal level.

Pre-filter trimmer on a 004 circuit doesn't just kill high end, relative to the mid dip around 400 Hz it dampens 4 dB. The balance changes greatly though.

Anyway, I'm not just making this 1 meg resistor up. See the picture attached. Steel string sultan I believe. Is it also present in gut shots of the original 005 or a Bludo?

[niusia]

Hi guys, first time posting here.

I was working on SSS#005 from AN pictures also, I was perplexed by 1Meg resistor feeding filter circuit, I was also perplexed by wet/dry mixing. These are results of my work. Please note, I have never heard or saw real SSS, AN or Bludotone copy of SSS and most importantly I have not built an amp based on my layout/schematic yet, so assume it has errors.

I would like to take this opportunity to thank all of you guys who contributed to this and many other forums, you guys literally changed my life.

[beasleybodyshop]

Wow! very clean layout! Thanks for sharing your work.

[rootz]

Interesting layout! Look a look more like AN's work I've seen. Just wondering: how sure are you it is a 33pF cap that is placed over the 1meg resistor preceding the filter section?

[niusia]

You're welcome. But the credit should go to Amplified Nation for posting many beautiful pics of his Singers and Sultans. I just did a monkey work.

One thing to note is that AN version is probably not exact part for part copy of SSS#005. I believe so because on different pictures you see different tweaks here and there. But I also believe that despite some tweaks and changes in layout (mostly in the mains filter caps to fit everything in smaller chassis) AN version of circuit itself is very close to Bludotone HPD. Now, how close HPD is to a real thing? This is the question way beyond my expertise.

Regarding the "33p capacitor" the value is correct. I am not 100% sure about connection though, however I can't think of anything else that would make more sense than this 33p bypassing 1Meg. I have also found a picture of what I believe is mhartman build of SSS#005 that shows this connection clearly. Again, I have no idea if it was copied from original, or just a honest guesswork, I attached the picture for the confirmation that an amp with such circuit was indeed built by a respectable builder.

There is still many unknowns, most importantly B+ voltages on different stages, because it can really alter the feel of the amp. I hope we can inspire some more knowledgeable folks to step in.

[rootz]

AN uses jumpers under the board I think. You can see the ends sticking through the eyelets on the top side in many other places and other amps. I can see those ends around the 33p cap too. Haven't seen the value of that cap before on a picture before. Now that one is confirmed, at least against an amp that is based on the HPD, which is supposedly based on the 005. I agree with you that there isn't any other logical connecting point for the 33p cap.

Filter section is interesting too. The 100k to ground after the 820k there seems to be missing (when comparing to the previous schematic). It only seems logical that the reverb return pot takes this purpose as part of the voltage divider now.

There doesn't seem to be a cap over the grid stopper of v1b, where you'd normally fins a 220k/470p combo on an 4th gen ODS. I can clearly see shrink tubing in some pictures and no extra solder connection for a cap. Still a bit off guess work though.

B+ voltages can be worked out from a Steel string sultan. Does AN use twin transformers? Resistor values can probably be seen in pictures. I've got a collection of those pictures as well, all published by AN on different media by the way.

[mojotom]

Very nice, thanks guys !

As a CF newbie is it ok to use those two “bias” network on each side of the B+ (22k/1N4007) joined together for bias supply ?

Also the Sultan seems to use choke vs resistors. That might be AN wanting more headroom from the amp.

What puzzled me a bit is the reverb supply. Usually dedicated from the screen supply something around 300-350V. Here the supply seems quite mower and part of the preamp supply (15K before 820R for V2 and 1k for V1).

PI values more in the line of the tweed values could make sense considering the tail (4k7) and feedback value.

[rootz]

Very nice, thanks guys !

As a CF newbie is it ok to use those two “bias” network on each side of the B+ (22k/1N4007) joined together for bias supply ?

Also the Sultan seems to use choke vs resistors. That might be AN wanting more headroom from the amp.

What puzzled me a bit is the reverb supply. Usually dedicated from the screen supply something around 300-350V. Here the supply seems quite mower and part of the preamp supply (15K before 820R for V2 and 1k for V1).

PI values more in the line of the tweed values could make sense considering the tail (4k7) and feedback value.

If I'm not mistaken AN takes the bias/negative CF voltage in the way you describe. Really something that should be checked against pictures. It should work anyway though. Resulting negative voltage is highly dependable of the tube used. This circuit above is designed for 12ax7 CF's. I can't say for sure you'll be left with enough bias voltage if you'd use a 12AT7 in the CF spot. I need to do more calculations and simulations to be sure.

I doubt the resistor vs choke is just for headroom purposes. Higher voltage resistors instead of a choke compress more, but also make the preamp voltage sag under heavy loading (that is, when the screens aren't on a separate node).

It looks like the power supply needs a bit of extra work. With a a double triode as a driver tube for the reverb you'd expect at least 3mA current draw there. That would drop the voltage available to V1 and V2 too more than we would expect. Might it be that the reverb driver is still on a separate node? I don't know for sure, can't really tell from the pictures. Any clear pictures from the supply around the can cap? That's where the questions (at least mine) are.

I believe feedback is taken from the 16 ohm tap on many pictures I saw. In one picture the PI cathode resistor is 470 ohm, like in e.g. a Marshall. In others it is hard to tell what the value is. In many I can't see a hint of yellow or violet, so it might not be 470 ohm in all AN amps.

[Stratoblaster]

Thank you for your great work and sharing beasley and niusia !

I have been spending some time trying to work out/come up with making the filter section work with the reverb mixing with fender style reverb setup.
Difficult to balance wet & dry signals , get enough top end response and gain levels all at same time.
I'm not necessarily trying to copy the HPD / #005 thing.

Yes some of those AN pics show different things going on to other pics.
As far as pots go, maybe this pic can help , if correct ?

Pots.jpg

Also I cannot see the 220k wet/dry mixing resistors on any pics Ive seen. Obviously one would be obscured under the board ,connecting directly to the input of the tube.
The other I do not see at all. The shielded wire connection between the hi filters and the top of the return reverb pot does not have a resistor at either end in any of the pics Ive seen.
I'm not sure where else it could be ?

Filters.jpg

Filters2.jpg

[erwin_ve]

They are in the heat shrink, you can see the shape of a resistor.

[Stratoblaster]

Thanks erwin, I realized that there were resistors under the heat shrink .I assumed one of them was the 270k resistor that connects to both hi & low filters.The other I assumed was the 12k to ground on the bass filter.
I need to take a closer look I guess.
Some of those AN pics are confusing. On one I see a 1meg resistor on the hi Filter. On another I can't see it. It might be obscured though.

About filters, my 2 cents worth.
On SSS#004 it appears HAD boosted the hi end frequencies with a 100k & 1000pf treble peaking network. Then the Hi filter is a cut like the low filters ,bleeding to ground.
As this method looks simpler, I always assumed something similar would be adopted for the SSS#005.
But it appears he may have returned back to a similar method to SSS#002. Making a variable treble peaking network on the Hi filter instead.