Ryan & Erwin's SSS #002 Madness

[Richard1001]

I measured some things and can confirm that with large signal (but smaller then you might think) and bigger cap selected on the high filter, the grid of V2 is drawing current. The tops of the sinewave are flattened on the positive half and unafected on the other half.
See the scope image, the 2400pf cap was selected.

IMG_20210102_170328.jpg

[Richard1001]

Today i measured the effect of the first feedbackloop with 1M and 100k.

Because the amount of feedback depends on the settings of the tonestack and volume control, i measured with the tone controls at full.
I did two measurements, one with the volume control closed (no feedback from the loop) and one with the volume control at full (maximum feedback from the loop)
I did this because even without the loop, a small feedback signal is present on the 100ohm resistor due to the ac current through the tube.
Because the feedbacksignal is so small and I had to set the scope at a very sensitive setting, there was some HF interference while measuring so the accuracy is +/- 0,5 mV.

The signal on the input was 30mV RMS @ 800 Hz

With the 1M resistor:
Volume control closed: 2mV RMS over the 100 ohm resistor, volume on full: 4mV RMS over the 100ohm resistor.
So with the tonestack and volume control at full the amount of feedback added by the loop is only 2mV.
Since (almost) nobody plays with everything on full, the effect of the loop is close to nothing, most of the feedback is coming from the 100 resistor by itself.

With the 100K resistor:
Volume control closed: the same 2mV RMS over 100 ohm, volume on full: 11mV RMS over the 100 ohm resistor.
So with everything on full, the amount of feedback added by the loop is 9 mV.

To put it into prespective I also measured the effect of the other negative feedback loops in the amp. Everytime with the same 30mV RMS signal on the grid of the triode with the 100 ohm feedback resistor on the cathode.

Reverb recovery (with 270K feedbackresistor): 30mV RMS on the grid, 2mV AC over the 100 ohm resistor with the return controll closed, 14mV AC with the return on full.
Second preamploop (470K feedback resistor): 30mV RMS on the grid of the second triode, 10mV RMS over the 100 ohm resistor. The 0.01 cap selected.

So it seems that only the first loop when using 1M does next to nothing where the other loop are. To me it seems logic that 1M is to high and should be lowered to 100K to give the loop any real function. Ofcourse there is no way to be sure what Dumble used except asking John Mayer to lend us his amp or ask the Japanese guy t use a higher resolution scanner to upload the guthot.

As for sound: both values sound great. No issues with both values.
With the 1M the amp is more Fender like and loose. With the 100K more like what i expected a SSS to react like. It is not exactly the same as compression but it is the only way to explain it. Also i noticed the tone controls are more ' active' with a lower resistor. Ofcourse this is all at bedroom levels since i am at home.

[Richard1001]

So yesterday I did more sound testing and measuring. To me the 100k feedback resistor is sounding the best. This becomes more obvious when overdriving the preamp. I did not do this before in the earlier sound test. The overdrive sound is very good with the volume turned up. (Even without any boost and pedals) much like a demo i have seen of a Sebago combo. I did not expect this because more negative feedback is commonly associated with bad clipping, but not in this case!
It ties the whole amp together. Making the tone controls and filters come more to live. Lastly, i noticed it keeps the amp more consistent sounding at low and high volume. Working better, i think i found a use for the mid switch. Putting the highfilter on the first position (3nF to ground) and engaging the switch results in a usable 'honky' mid sound with enough clarity.

Next i turned my attention to the reverb feedback resistor. On the SSS 001 it is cleary 100K but in my amp and with the reverbtank i use (accutronics usa 2006, 9ab2c1b) there was not enough reverb. Because the SSS 002 has an additional feedbackloop on the driver that the 001 does not have, to me a larger feedback resistor to make up for this additional loop seems logic. On the japanese schematic it says 270K but this dit not sound exactly right to me. On some notes the reverb was ringing longer than i like and this got in the way of playing. Even at lower reverb settings. Testing different resistors i found that the feedbackloop resistor controls the amount of reverb available but also does something to control the decay of the reverb. I wanted a good balance between enough reverb and a musical sound so that even with a lot of reverb, it doesn't get in the way of my playing. After testing all options to me a 180K resistor sounds the most musical and has the best balance between a good strong reverb sound level, without any long decay of some notes ringing that get in the way. Looking at this resistor on the gutshot of the SSS 002, there is no sure way to tell what is in there. It might be 270, 180 or 100K. Maybe this resistor was adjusted to the reverb tank used, because in my experience the output signal and decay can varry somewhat on accutronics tanks of the same type. Who knows?

[erwin_ve]

So yesterday I did more sound testing and measuring. To me the 100k feedback resistor is sounding the best. This becomes more obvious when overdriving the preamp. I did not do this before in the earlier sound test. The overdrive sound is very good with the volume turned up. (Even without any boost and pedals) much like a demo i have seen of a Sebago combo. I did not expect this because more negative feedback is commonly associated with bad clipping, but not in this case!
It ties the whole amp together. Making the tone controls and filters come more to live. Lastly, i noticed it keeps the amp more consistent sounding at low and high volume. Working better, i think i found a use for the mid switch. Putting the highfilter on the first position (3nF to ground) and engaging the switch results in a usable 'honky' mid sound with enough clarity.

Next i turned my attention to the reverb feedback resistor. On the SSS 001 it is cleary 100K but in my amp and with the reverbtank i use (accutronics usa 2006, 9ab2c1b) there was not enough reverb. Because the SSS 002 has an additional feedbackloop on the driver that the 001 does not have, to me a larger feedback resistor to make up for this additional loop seems logic. On the japanese schematic it says 270K but this dit not sound exactly right to me. On some notes the reverb was ringing longer than i like and this got in the way of playing. Even at lower reverb settings. Testing different resistors i found that the feedbackloop resistor controls the amount of reverb available but also does something to control the decay of the reverb. I wanted a good balance between enough reverb and a musical sound so that even with a lot of reverb, it doesn't get in the way of my playing. After testing all options to me a 180K resistor sounds the most musical and has the best balance between a good strong reverb sound level, without any long decay of some notes ringing that get in the way. Looking at this resistor on the gutshot of the SSS 002, there is no sure way to tell what is in there. It might be 270, 180 or 100K. Maybe this resistor was adjusted to the reverb tank used, because in my experience the output signal and decay can varry somewhat on accutronics tanks of the same type. Who knows?

Hi Richard, thanks for sharing you observations. I'll try the v1 Feedbackloop tomorrow in my own SSS002.
The experience it is sounding better when overdriving the amp is no suprise to me; the ODS #124 and #102 with the NFB loop (although is different) sound more smooth than without.

I also had some experimenting with different reverb tanks; I use the 4ab3c1b and different vendors of the same model had different measuring coils . The TAD was the lowest in output, the accutronics the highest. The accutronics is in my build now hence your conclusion the reverb NFB is adjusted to the tank makes perfectly sense!

Erwin

[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?

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).

[Richard1001]

Hi Erwin,

There are a quite a few mistakes in this schematic, and staying in there consistently.
For instance, take a look at the deepswitch, filters or master volume. It could get confusing for someone who wants to build the amp.

[erwin_ve]

Hi Erwin,

There are a quite a few mistakes in this schematic, and staying in there consistently.
For instance, take a look at the deepswitch, filters or master volume. It could get confusing for someone who wants to build the amp.

Hi Richard, Thanks, Ill take it offline, when updated Ill post it again.

[erwin_ve]

The schematic is edited; pls check for errors.

[Bombacaototal]

https://raw.githubusercontent.com/colga ... Layout.png

I have linked above the latest available layout for ease of access in regards to a question.

As far as PT wiring goes, is there any practical difference between having the brown plus black wire (0V) in the mains ON/OFF (non fused side) and the brown/White plus black/white (120V) on the receptacle (N) or having these reversed with the brown plus black on the receptacle (N) and the other two on the mains ON/OFF?

In regards to the bias tap on the PT (50V) this is not used as the bias circuit needs to generate the negative voltage instead, correct

In regards to the OT is there anyway to check if primaries are not reversed before start up?

[erwin_ve]

https://raw.githubusercontent.com/colga ... Layout.png

I have linked above the latest available layout for ease of access in regards to a question.

As far as PT wiring goes, is there any practical difference between having the brown plus black wire (0V) in the mains ON/OFF (non fused side) and the brown/White plus black/white (120V) on the receptacle (N) or having these reversed with the brown plus black on the receptacle (N) and the other two on the mains ON/OFF?

In regards to the bias tap on the PT (50V) this is not used as the bias circuit needs to generate the negative voltage instead, correct

In regards to the OT is there anyway to check if primaries are not reversed before start up?

1. In EU there isnt any difference, the wall voltage pins are not always wired to a specific pin. Although when studying for EE(in previous century, Im old) there were specific rules, but hardware doesnt indicate that. Other countries might have specific wall voltage wiring concerning the hot and neutral wire.

2. The bias tap is not used as it is derived from one of the the HT wire(reverse diode). Negative or positve voltage is just a matter of how the diode is wired.

3. Ot primaries; Brown wire most far away from the PI, Blue wire closest to the PI is usually correct.
Something Martin Manning posted; "Connect the primary's CT the the hot side of the secondary, put some AC voltage on the secondary, and then measure voltage from the secondary ground to each plate lead. The lead with the higher voltage is the in-phase side"

[martin manning]

3. Ot primaries; Brown wire most far away from the PI, Blue wire closest to the PI is usually correct.

The in-phase primary lead goes to the side that is driven by the plate of the PI's input triode.

Something Martin Manning posted; "Connect the primary's CT the the hot side of the secondary, put some AC voltage on the secondary, and then measure voltage from the secondary ground to each plate lead. The lead with the higher voltage is the in-phase side"

Yes! And if the OT schematic has phase markings, you can identify the in-phase lead from that, but I always check to make sure all the lead colors are correct.

[Bombacaototal]

Many thanks Erwin and Martin!

Good to know in regards to the PT wiring. I am glad there is no difference.

Noted in regards to the bias

And thanks for the tip on the OT primary

Last but not least, the layout has the same ground for mains and PT (incl bias). I read that is advisable to have a dedicated separate ground for the mains. I at least have done this in all of my other amps. Is there any real benefit in doing so, or is it fine to aggregate the grounds as per layout

[erwin_ve]

Many thanks Erwin and Martin!

Good to know in regards to the PT wiring. I am glad there is no difference.

Noted in regards to the bias

And thanks for the tip on the OT primary

Last but not least, the layout has the same ground for mains and PT (incl bias). I read that is advisable to have a dedicated separate ground for the mains. I at least have done this in all of my other amps. Is there any real benefit in doing so, or is it fine to aggregate the grounds as per layout

Ground for mains, seperatly, as close as possible to the mains connector is mandatory in the EU. Make sure not to use the screw attached to the mains connector!!

[Bombacaototal]

Makes sense, many thanks Erwin!

[Bombacaototal]

I've been a bit on a transformer nightmare ride which I want to share and also have a few questions for the members.

I was an avid user of Classic Tone transformers, and always prefered the vintage spect 320-0-320 HV taps (versus the 345-0-345 present as an added option in a few of their transformers). With mains voltage of approx. 240V and using the PT 240V tap, I usually get around 455V in my Twin and Bassman 6L6 amps.

With the advent of CT going out of production I decided to custom order a set from a manufacturer in the Netherlands (won't name names unless someone specifically asks). I ordered a paper wound vintage Fender Twin replica with 120, 230, 240 mains and vintage spec 320-0-320.

I received the transformer poorly packadged and attached is the situation I was comfronted with. Basically the transformers were sent unwrapped in a polysterine enclosure, which reacted to the materials used and basically glue onto the transformer. The builder offered me to send them to a business partner in the UK who properly cleaned them, to a reasonably good state and gave me a small refund as a gesture of good will. At least this issue was sorted, and the manufacturer was helpful in sorting it.

I just finished building the amp and I am getting only between 415 and 420V on the 6L6 plates. I contacted the manufacturer who wrote the following

320VAC rectified with solidstate diodes gives 450V. This 450V will when you put it from standby to play position drop to 423V because the four powertubes draw 30mA each. The 423V will drop about 5V further down through the Outputtransformers Ohms resistance to about 418V. So your 417V measurement is very accurate.

It is simple math. 320 x1,414 = 450V without load. 450 x 0,94 = 423V under load in Class AB. 4-5V losses over the primairy winding of the OT. These are laws that you cannot change. They will apply on every tube amplifier.

I went on to ask why those old fender transformers would usually read between 440V and 460V and have 320-0-320? Below is his answer

What can be is that they give a spec for 320V AC under load. That means that their spec unloaded must be around 340V AC

So my question is, how were the actual spec on those old Fender transformers? Is it 320V loaded or unloaded? Is there anything that can be done on my case? Is there any obvious shortcomings of such low plate voltage besides lower than 100W power?