Ryan & Erwin's SSS #002 Madness

[Richard1001]

I allready made a comment on that. Ryan and Erwin both reported the amp could use more low end. Ryan suggests to use larger coupling caps in certain positions (.022 instead of .01) They use the 1uF de-couplingcap on both mixing stages. My opinion is that the 5uf de-couplingcap shoud be on the mixing stage of the direct signal and the 1uF de-coupling cap on the reverb mixing stage. To me, this makes sense.

If you calculate the cut off frequency with a 1 uF cap decoupling the 1k resistor, you get -3 db around the frequency of the low A string. It makes no sence to reduce low frequency gain for the direct signal at this point in the schematic. It does however make sence to have less gain for the low frequencies at the reverb side to keep the reverb clear without to much (the for this effect unwanted) low frecuency's. My opinion is that the person who drew the schematic used the right values, but accidently turned them around.

In my amp i used 5uf on the direct signal mixer and 1uf on the reverb side. I did not use and dit not need to change to the bigger couplingcaps Ryan suggests to get a decent bass responce. I just use the smaller coupling cap values as in the original japanese schematic and have a very strong bass response.

[martin manning]

My opinion is that the 5uf de-couplingcap shoud be on the mixing stage of the direct signal and the 1uF de-coupling cap on the reverb mixing stage. To me, this makes sense. ... My opinion is that the person who drew the schematic used the right values, but accidently turned them around.

This is exactly what I was thinking too. Going from 1 to 5u is worth +2dB on the bass side.

[hermannbjorgvin]

What are the mounting nut plates required for this chassis layout? I’ve tried the recommended 10-32 K1000-3 and MK1000-3 from Aircraft Spruce but neither fit the chassis mounting.

I had no idea these were a thing... I just went to the hardware store and bought screws and nuts...

Only my second amp tho, last one was a Champ.

Is it possible to use the Fender chassis screws with the right screw anchor? The Fender screws are a little short to make it through the chassis

[rccolgan]

Sorry to leave you guys high and dry for a while. Neck deep "distance learning" my 7 year old daughter with not much time left for this stuff until she's back full time. Full chaos mode currently.

I updated the LT spice schematics and pdf. The simulations look much better now too. My suggestion is to download the LT spice files and add corrections & post them here and I'll integrate the changes. I took extra effort to make sure the dependencies are all self contained in the download so you should be able to just download & run in LTSpice.
https://github.com/colganr/LTSpiceCircu ... er%20sn002

As for the anchor bolts, I did make a mistake and copied/pasted from a CAD file that had wrong dimensions for the outside screws. My suggestion is to get the airplane nut and simply drill the outer holes you need for the nut you bought. Shouldn't take more than a few seconds to modify. Personally, I use a 4" long bolt from Ace Hardware and a Keps nut.

Great conversation on the bias alternative and how to introduce more bass into the circuit. This morning, I put mine back to "stock" .01uf coupling caps per the Japanese schematic as well as the 5uf bypass cap on v3. Playing through my Dumble style 2x12 with Eminence EM12N the amp sounds really well balanced without any cap swaps. I should note that the bass comes in and evens out when the amp is cranked to stage volume. (thank goodness for load boxes)

-Ryan

[hermannbjorgvin]

My suggestion is to get the airplane nut and simply drill the outer holes you need for the nut you bought. Shouldn't take more than a few seconds to modify.

I'm going to get a few SSS #002 specific nut plates machined in case anyone is interested. Basically a small aluminum block with two 6-32 threaded holes and a 8-32 threaded hole with a small countersink to easily receive a fender chassis screw.

I'll put anyone who's interested down for a set of four, I'll get a quote as soon as I draw it up and talk to the machining shop, it should not be too expensive.

Anyone interested just PM me, I'll give you a chance to back out when I get the final cost+shipping

[martin manning]

Sorry to leave you guys high and dry for a while. Neck deep "distance learning" my 7 year old daughter with not much time left for this stuff until she's back full time. Full chaos mode currently.

No worries, these are trying times!

I updated the LT spice schematics and pdf. The simulations look much better now too. My suggestion is to download the LT spice files and add corrections & post them here and I'll integrate the changes. I took extra effort to make sure the dependencies are all self contained in the download so you should be able to just download & run in LTSpice.

Great, Thanks!

Great conversation on the bias alternative and how to introduce more bass into the circuit. This morning, I put mine back to "stock" .01uf coupling caps per the Japanese schematic as well as the 5uf bypass cap on v3. Playing through my Dumble style 2x12 with Eminence EM12N the amp sounds really well balanced without any cap swaps. I should note that the bass comes in and evens out when the amp is cranked to stage volume. (thank goodness for load boxes).

Good to know. There are other options for the bias adjust, such as a level and balance arrangement. Conversion to the bipolar driver supply is easy; the extra components will fit on a small board along with the HV rectifiers, and from there it's just a few resistor value changes in the driver section.

[hermannbjorgvin]

Can someone explain 5uf on V3 to someone who's new to amp building? Is it adding bass? Is it more likely to be the value of the original?

Also while I'm asking what's the consensus on the zener diode? Is it required to be able to use the standby for longer periods? Not that using a standby on an amp for long periods is good anyway.

And final question, does anyone know of a footswitch relay board that has LED indicator lights for the footswitch? Is it correct of me to assume those would be powered by the relay board?

[Richard1001]

There are some things you must know to understand wat a cathode bypass cap does and how it effects the gain of a cathode base amplifier stage.

1. A cathode base amplifier stage will amplify the diffrence in voltage between the grid and cathode.
2. The impedance (ac resistance) of a capacitor depends on the frequency of the signal and value of the cap.
3. A current flowing trough a resistor will result in a voltage over the resistor.

Without a cathode bypass cap, the ac current trough the tube resuts in an ac voltage over the anode resistor (this is your amplified signal) and an ac voltage over the cathode resistor. The ac voltage over the cathode resistor will be in phase with the ac voltage on the grid. So the diffrence in voltage between the grid voltage and cathode voltage (the voltage amplified by the tube) will be smaller than the voltage on the grid itself. This is a form of negative feedback and the gain of the amplifier stage is greatly reduced

To eliminate this negative feedback and get more gain from the tube, you want to keep the voltage at the cathode the same and not have it move up and down in phase with the signal on the grid. In order to do this, you need to tie the cathode to ground for ac. This is what the purpose of the cathode bypass cap is. It is an ac coupling to ground in order to keep the cathode voltage at the same level. With the cathode voltage at the same level, no longer moving up and down with the grid voltage, the diffrence in voltage between the grid and cathode is the same as the grid voltage itself and the gain of the tube stage wil be (much) higher. There is no more negative feedback.

Since the impedance of a cap depends on the frequency of the ac signal and value of the cap, it is essential you select the right value cap in order to tie the cathode to ground and bypass the cathodecap sufficiently for the lowest frequency you want to amplify without the negative feedback from the cathode resistor.
The smaller the value of the cap, the higher the impedance of the cap will be for lower frequency's. As the frequency gets higher, the impedence of a cap will get lower. So a smaller cap will result in higher impedence (or ac resistance) for low frequency's and will have a lower impedence (or ac resistance) the higher the frequency. This means the tube stage will have frequency dependant negative feedback (and thus gain), because for the lower frequency's the cathode voltage can move up and down with the signal on the grid (cap has higher ac resistance, allowing negative feedback), and the higher the frequency's gets the lower the impedance of the cathode cap gets, stopping the cathode voltage moving up and down with the grid.

Designing an amp, you can use this to determin the frequency responce of a tube stage. Sometimes you want to get rid of low end. For instance to get a tight overdrive sound without muddy lows. An example of this is the 0.68uF cathode bypass cap on the first stage of a Marshall amp. In old fenders there is a very big cathode bypass cap on the first stage. This big value was not so much for the sound of the amp, but to get rid of the induced 50Hz (or 60 Hz) hum from the heaters to the cathode by shorting it to ground.

Anyway, in the mixingstage V3 of the SSS, it makes no sence to use a smaller cathode bypass cap then the other direct signal stages before that. There is no need to get rid of the low end. By the time the signal gets to this point, it allready passed the low filterswitch, deepswitch and baspot to controll the amount of low end.

So yes, i firmly believe 5uF on V3a is the right and original correct value.

The reverse diode from B+ (screencap) to ground is required in my opinion to protect the filtercaps in standby. (Reverse diode meaning anode to ground, cathode to B+, a regular 1N4007 will do fine) The zener was ment to lower the voltage on the anodes of V7 and is not required.

You can just use leds switching in series with the relays on a stable 12 volt (or so) supply. Use relays with 200mW coils and plain 20mA leds. The coils of the relays will act as dropping resistors for the leds. I have done this and use 3 relays with a 3 switches and 3 led footswitch and a 4 conductor wire. Works great. Best to use the positive as common for all the relays, and the other side of the coil switching to ground (with series led or without, both will work) This way you can never short the relay supply.

[hermannbjorgvin]

Thanks Richard for the detailed explanation!

I ordered a breakaway 3 DPDT relay board from Granger https://grangeramp.com/product/3-relay- ... d-ver-2-2/

Looks like it has all the features I need for the LED lights in series as well as switching. I'm gonna see if there's space to mount this on the side of the chassis for a Reverb and FET footswitch, if it doesn't fit it'll just be a reverb switch like the original then.

[martin manning]

I ordered a breakaway 3 DPDT relay board from Granger https://grangeramp.com/product/3-relay- ... d-ver-2-2/

You'll need a 5VAC source for that.

[hermannbjorgvin]

I ordered a breakaway 3 DPDT relay board from Granger https://grangeramp.com/product/3-relay- ... d-ver-2-2/

You'll need a 5VAC source for that.

Hmm it says 6.3VAC from the filaments is possible with a floating ground. Did I misunderstand something about the center tap there?

[martin manning]

Granger is saying you can use the filament supply (which will produce ~7.5VDC for the relays), and ground either the filament CT or the relay power supply negative (not both). They claim it's quiet if you do the latter, so I would go for that so that there is no danger of accidentally grounding the relay supply. The relays should tolerate the higher voltage, but check the data sheet. If the LEDs are in series, the diode voltage drop will take care of that.

[Richard1001]

On the 3n cap connection on the high filter: In the plots below the two options are shown. In the first, isolating the 3n produces a huge treble boost in all the other positions, which suggests that it is an error. Connecting the 3n to the common terminal of the other caps and then to ground produces a much more reasonable result. The third plot is the revised filter I developed, which evens out the increments and reduces interaction.

Today i did some experimenting and found that following the schematic that Ryan drew of the high filter with the 3nF cap connected to the common terminal of the other caps, there is a huge boost in volume when this setting is selected. This makes sence. With this wiring selecting only the 3nF cap wil also short the 820K resistor resulting in a boost of volume. The strange thing is that I would expect this to also show up in a simulation.

On the demonstration video of the real Dumbleland special it is clear that there is a highcut setting on the high filter switch. When selected in the demonstration it does not result in a volume boost, just a treble cut. I think this can only mean that the 3nF cap to ground is indeed a isolated setting and that it is not connected to the common terminal of the other caps.

What are your thoughts Martin?

[martin manning]

What are your thoughts Martin?

In the hand drawn schematic there is an input network followed by a low-pass branch and a high-pass branch, which are then merged at the output, where there is something like 70 to 100k to ground. The low-pass branch (the high filter) has a series of caps placed across an 820k resistor, and then one odd setting where there is no cap across the 820k, and a 3n to ground after the 820k. That last setting will make a low-pass filter with a big treble cut. The question is, is that what was intended, or is there some connection missing from the drawing?

For sure the lowest (CCW) setting on the Carter's Vintage Dumbleland Special demo has a large treble cut (https://www.youtube.com/watch?v=zthE05CGFvY just at 1:30), and you do see that in the plot I posted above with the 3n isolated. The wide-band volume boost in all the other settings indicated in the plot doesn't seem to show up in the demo though, the treble just drops out. That setting could just be a small cap, but in any case it seems to be out of order in the frequency sweep of the filter. It's possible that the filter circuit in the demo amp is not the same as the one in the #002 SSS, or maybe the cap on that setting was a later modification, applied without rebuilding the whole filter.

Referring again to the hand-drawn schematic, its clear that the connection of the 3n to the common capacitor bus is not there, and has been added to redrawn versions based on speculation or some other source of information. That arrangement seems to make the frequency response plot look better, and the resulting short over the 820k does not change the response much. It looks like a large cap, but it isn't much different from the setting where the 5100p value is switched in. That seems redundant.

IMO, the high filter makes most sense with the 3n connected to the center pole of the switch, so the filter is essentially an 820k followed by a 3n to ground (basic LP filter), with various cap values available to bypass the 820k. That's easy to try out by jumping the non-ground side of the 3n to the switch pole. I think that will more closely match the Dumbleland video, with a treble cut on the 3n cap position.

PS: The cap values on the hand-drawn schematic are a bit confusing, as they appear to be marked 150, 330, 1000, .01, 2400, 5100. The .01 is really out of place. Others have adjusted the values to roughly multiples of two, which seems appropriate.

[Richard1001]

Interesting, I can't find the 'no cap over the 820k' position in rhe japanese schematic.

As for shorting the 820k resistor: maybe i was not clear, but with the 3nF cap on the common terminal (white wire) it is the switch itself that shorts the 820k resistor causing the volume boost. If the switch would select the 3nF cap to ground it shorts the green and white wire witch are over the resistor. This is a fault for sure in my opinion. (White and green color code on the japanese swchematic)

As for the 10nF cap being out of place: there is a simple answer for that. On the builds here on ampgarage, most people use a 7 position switch. I believe Dumble origialy did not and used 11 and 12 position switches witch can be blocked anywere between 2 and the maximal number of positions of the switch. Now, if you use a 11 position switch for the high filter like the one drawn in the japanese schematic and block it for 7 positions, it leaves some unused terminals you can use for the common capacitor buss (again, like drawn in the japanese schematic) Since the 10nF cap is to big to place on the outside of the switch (where the common terminal and terminal to connect the cap are close together) it is placed in the middle to the common terminal and connected with an angle under the other caps to the correct terminal so it sits nice on the switch.

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

The drawing of the switch is actualy drawn as a layout in my opinion, not a schematic. I implemented this and posted a picture earlyer of the swtches and wiring i used.