Ryan & Erwin's SSS #002 Madness

[martin manning]

On SSS front panels, (and also in the Dumbleland demo video linked above), the filter switches have seven positions with the end points 180 degrees apart, which corresponds to seven of twelve equally-spaced positions used. The hand drawn schematic also shows twelve positions, and seven in use. All good there. I'm not following you on the .01 cap, though. I see six caps all tied together at one end, and six values listed with the .01 in between 1000 and 2400. The 3000p going to ground is on the first switch position, and it is not connected to the other caps. When that position is selected, there is no cap connected across the 820k. I've marked up the drawing of the filter section below to show how I'm seeing it.

[Richard1001]

On SSS front panels, (and also in the Dumbleland demo video linked above), the filter switches have seven positions with the end points 180 degrees apart, which corresponds to seven of twelve equally-spaced positions used. The hand drawn schematic also shows twelve positions, and seven in use. All good there. I'm not following you on the .01 cap, though. I see six caps all tied together at one end, and six values listed with the .01 in between 1000 and 2400. The 3000p going to ground is on the first switch position, and it is not connected to the other caps. When that position is selected, there is no cap connected across the 820k. I've marked up the drawing of the filter section below to show how I'm seeing it.

I thought you were revering to two seperate positions on the switch. A 3nF to ground and a 'nothing over the 820k' position, but now i understand that you mean that this happens both in the same switch position and i agree that's correct.

On the schematic you posted it is clear that the 820k will be shorted by the switch if the 3nF would be conected at the common terminal of the other caps. So to me this would be wrong and is no option because of the huge volume boost it creates. So the 3nF either has to be an isolated position to ground as drawn in the japanese schematic, or it should be connected to the switching contact of the rotary switch like you suggested. Both will work. Whats left to determin wich way is the way Dumble did it, is the demonstration video of the Dumbleland special. This shows a clear treble cut in the first position. Of the two options left only with the 3nF isolated to ground this treble cut is achieved. So to me this is the way it should be wired to get it the same as the original SSS 002. To me it's all about figuring out how the original Dumble SSS 002 is wired.

About the position of the 10nFcap : The 10nF cap physicly sits between the 1000 and 2400 pF cap on the back of the switch. This can also be clearly seen in the gutshot of the amp. So it is my guess that it was accedently drawn connected in the middle (like what you would see looking at the back of the switch) when in fact the leg of the cap is connected with an angle under the other caps to position 7. I think i read about this mistake in the drawing somewhere before on this forum and it made perfect sence to me. When i wired the switch it became clear why. The 10nF would not fit nicely on the switch otherwise.

About the plots:
I have been looking at the plots again and was wondering about something: Wouldn't it be correct to say that the response curves of the low and high switch should be added together to see what realy happens? In the plots they are drawn as separate curves, but adding the curves of the high and low filter together in all diffrent positions realy shows what happens. To explain what i mean, think about a crossover filter: On a crossover there is a curve for the low filter and a curve for the high filter. Adding both together results in the curve that shows the actual frequency response. (In the case of a crossover the two curves ad up to a straight line, more ore less)

[Richard1001]

About the plots, never mind, I didn't read them right. Looked at them again and saw that the curves are allready added together.

[martin manning]

On the schematic you posted it is clear that the 820k will be shorted by the switch if the 3nF would be conected at the common terminal of the other caps. So to me this would be wrong and is no option because of the huge volume boost it creates. So the 3nF either has to be an isolated position to ground as drawn in the japanese schematic, or it should be connected to the switching contact of the rotary switch like you suggested. Both will work. Whats left to determin wich way is the way Dumble did it, is the demonstration video of the Dumbleland special. This shows a clear treble cut in the first position. Of the two options left only with the 3nF isolated to ground this treble cut is achieved. So to me this is the way it should be wired to get it the same as the original SSS 002. To me it's all about figuring out how the original Dumble SSS 002 is wired.

All three possibilities work, but the one that looks most unlikely to me is having the 3n isolated, where I see a huge volume jump across the upper frequencies, except for position one. Connecting the 3n cap to the capacitor common bus (as others here have done) works, and doesn't show a large volume jump in the simulation, but I don't favor that option either. The essence of a low pass filter is a resistor followed by a cap to ground, and that is what you get by connecting the 3n to the switch pole. That arrangement shows a substantial treble cut in position one, and the overall response looks a lot like the demo video sounds. I'm betting that is how the original is wired.

[Richard1001]

All three possibilities work, but the one that looks most unlikely to me is having the 3n isolated, where I see a huge volume jump across the upper frequencies, except for position one. Connecting the 3n cap to the capacitor common bus (as others here have done) works, and doesn't show a large volume jump in the simulation, but I don't favor that option either.(...)

My conclusion is that there must be a diffrence between the simulations and the response of the actual amp. You see things on the simulations I can't hear, and I hear things that don't show up on the simulations.

With the scenario of the 3nF conected to the common bus:

If there is no volume boost on the simulations between position 2 ( a voltage divider between the input and output of the filter consisting of 150p//820k into 100k) and position 1 seleciting only the 3nF cap (wich results in a hard wired connection between the input and output of the filter into 100k), there must be something wrong with the simulations.

It is simply impossible to have the same output volume between these two settings.

IMG_20200921_190639.jpg

[martin manning]

The 1n in front of the cap array keeps the level down. Shorting the 820k is like putting a large cap across it, not too much different from the 10n being switched in.

[Richard1001]

The 1n in front of the cap array keeps the level down. Shorting the 820k is like putting a large cap across it, not too much different from the 10n being switched in.

I know the 10nF would have a very similar effect, that to me is just another reason this wiring wouldn't be right.

And the 1nF won't keep the level down. The whole point of the filter is that the level of the high frequency's go up when selecting a bigger cap. And when this happens the overall volume of the amp goes up as well. In the demo of the dumbleland special this happens gradualy all the way up. Wiring the 3nF to the other caps would result in a huge step at once between position 1 and 2 and this huge step is not present in the demo of the real thing.

And because not only the level but also the total capacitance (1nF in series with the selected cap) moves up, this would result in a gradualy widening frequency band. (Like selecting a different treble cap)
That to me would be the whole point of this filter.
If the purpose of the filter would be to move the same high frequency band up and down in even steps, Dumble just could have used 1 single cap and a resistor network.

So, I think it's time to say we agree to disagree on this one

[martin manning]

Maybe I misunderstood. I agree that a large jump will occur (on the treble side) from position 1 to position 2 if the 3n is tied to the cap common (making position 1 look like position 7), and that does not occur in the demo. But, isolating the 3n to ground on position 1 makes the treble response in all the other positions jump much higher, as shown in the plot above.

[rccolgan]

The order is 3nF, 150, 330, 1000, 2400, 5100, 10nF making 7 positions.

The drawing of the switch is actually drawn as a layout in my opinion, not a schematic. I implemented this and posted a picture earlier of the switches and wiring I used.

Hi Richard - Amazing contributions!!!!! I cannot thank you enough. You caught one of my mistakes on the schematic On the layout, I did ok, I think (attaching picture). I copied Ceriatone's SSS filter layout and copied in #002's values and glad I did vs following the hand-drawn Japanese layout of the filter section. This is also why the rotation clockwise cuts vs opens up as well. Rotation aside, I feel like this is in-line with the Dumbleland video?

I caught up with Erwin this morning as well so we should see some minor changes in my main photo album. Mainly with the protection diode and PCB value changes. https://photos.app.goo.gl/W8zAvFVRCSsRoQgs7

[Michiel ODB]

Hi all,

Almost done with my version of the SSS! Made chassis & boards by myself and the pre amp is already sounding beautifull!

However, some problems come up in the power stage, I think it might have to do something with my PI. Maybe some guru's in here can assist me with my problem, cause I'm almost out of options , (First time on this forum also, so if this isn't the right place to ask for help, please let me know! )
I will try to explain my problem as clear as possible.

The sound I'm getting out of the amp is very weak at full volume, and way too distorted. When I put my scope on the output of the power tubes, I can see the original signal, but also a lot of ripple and unwanted behaviour. I'm guessing al this extra 'noise' Is cancelling out the actual sound in the output trannie ... (not really sure of noise is the best way to describe it btw)

When I add a clean sinus signal on the input of the PI (say 440Hz, about the same magnitude as I would normally have on the end of the pre amp), I can trace it with my scope, all the way up to pin 2 of the first valve (V5, in Ryan and Erwin's github schematic). However, on pin 1, the signal is "wobbles" all over my scope screen. I Think it might have to do something with a ripple on the supply voltage that reaches the trimpot (D, in the schematic), but than I would expect the same behaviour on the other first pins in the amp, but the preamp just sounds fine.

Some ideas?
Already rebuild the PI and power section, have the right voltages on all tubes, and the output trannie seems to be fine.
(note: the picture from above is before I re-did some components, so some wiring might look different in the other pic)

[erwin_ve]

The power Section:
Looks fine, but I’ll suggest an alternative to the bias adjustment shown in the layout that does not require a 2W pot: Replace the 130k resistor going to ground from the junction of the 1M grid resistors with a standard 150k pot or trimmer (the element taking the place of the 130k), then connect its wiper to ground with a 220k resistor. This will provide a fail-safe adjustment from 89k to 150k. The 2W pot can then be replaced with a fixed resistor to get a suitable voltage range.

Martin you posted this a while ago; thanks for your suggestion. Im updating the PCB with this;
I this how you figured it on the driver side of the bias?

[erwin_ve]

Hi all,

Almost done with my version of the SSS! Made chassis & boards by myself and the pre amp is already sounding beautifull!

However, some problems come up in the power stage, I think it might have to do something with my PI. Maybe some guru's in here can assist me with my problem, cause I'm almost out of options , (First time on this forum also, so if this isn't the right place to ask for help, please let me know! )
I will try to explain my problem as clear as possible.

The sound I'm getting out of the amp is very weak at full volume, and way too distorted. When I put my scope on the output of the power tubes, I can see the original signal, but also a lot of ripple and unwanted behaviour. I'm guessing al this extra 'noise' Is cancelling out the actual sound in the output trannie ... (not really sure of noise is the best way to describe it btw)

When I add a clean sinus signal on the input of the PI (say 440Hz, about the same magnitude as I would normally have on the end of the pre amp), I can trace it with my scope, all the way up to pin 2 of the first valve (V5, in Ryan and Erwin's github schematic). However, on pin 1, the signal is "wobbles" all over my scope screen. I Think it might have to do something with a ripple on the supply voltage that reaches the trimpot (D, in the schematic), but than I would expect the same behaviour on the other first pins in the amp, but the preamp just sounds fine.

Some ideas?
Already rebuild the PI and power section, have the right voltages on all tubes, and the output trannie seems to be fine.
(note: the picture from above is before I re-did some components, so some wiring might look different in the other pic)

Your scope pic suggests the signal is getting through but fluctuating on on some sort of low freq ripple?
Cap or tube related perhaps?

[martin manning]

The power Section:
Looks fine, but I’ll suggest an alternative to the bias adjustment shown in the layout that does not require a 2W pot: Replace the 130k resistor going to ground from the junction of the 1M grid resistors with a standard 150k pot or trimmer (the element taking the place of the 130k), then connect its wiper to ground with a 220k resistor. This will provide a fail-safe adjustment from 89k to 150k. The 2W pot can then be replaced with a fixed resistor to get a suitable voltage range.

Martin you posted this a while ago; thanks for your suggestion. Im updating the PCB with this;
I this how you figured it on the driver side of the bias?

Yes, that was my proposal, which was aimed at easily modifying an existing amp. If you're redoing the PCB, I think I would keep the pads for a board-mounted resistor (the 130k), but leave the ground path open, to be connected by a jumper (fixed bias) or by a remote pot wired as a variable resistor (adjustable bias). The resistor would be 130k for fixed, 118k + 25k pot for adjustable, and this would also work using my proposed +/-170V bipolar driver supply with 24k + 10k pot (a standard Fender chassis-mounted trim pot).

[martin manning]

As mentioned above, another source of step filter information is the C-tone SSS, which I believe is based on #004. The published layout and a schematic (presumably drawn by someone with access to the hardware) shows that the cap and resistor arrays are very similar to the #002 drawing, including the isolated cap to ground on the high switch position 7. So, the isolated cap (which is reduced in size on the C-tone) seems likely to be correct as shown on the #002 drawing.

There is a significant change to the input network on the C-tone, which spreads the effect of the high selector over a broader frequency range and makes the HF cut of the isolated cap seem a little less out of place (along with the reduced value). I also note that the inductor (a large open-frame iron core unit mounted on the top of the chassis, shown as two 0.39H in series on the schematic) seems to have very limited effect, and only at ~20 kHz and above. It could be deleted.

If this circuit is from #004, it seems likely that it's an evolution of #002. Overall, the revised input network seems to have converted the high filter into much more of a mid control, impacting frequencies of a couple of hundred Hz and above at its highest setting. The C-tone documentation says: "NOTE – turning the HIGH filter all the way counter clockwise (to the left) will bypass the SSS’s 7-way rotary filter controls." It's not a bypass, as the low filter is still effective, but it does open it up from the low-mid range on up.

[Vertigo]

Ryan/Erwin

I noticed that one of the 22uF caps on the power board has been replaced with a dual 33uF. How critical is this change?