trinityamps.com

Hallo, since I built the amp I have made several modifications, tested, changed back a few ones that were not good... and learned a lot in the process.
Today I thought I would investigate why the precense control almost was non functional on the amp. Perhaps if parts of the schematic comes from an original 100W design, but this version is now 24W with two 6V6, then perhaps the voltage that goes into the negative feedback is too low with the resistors, pot and cap in the negative feedback circuit (4,7kohm, 390 ohm, 2k pot an 1uF cap). In addition the amp was always sounding a bit harsh with distortion up, especially compared to some of my other amps, which suprised me.

So I moved the negative feed back signal from the 4 ohm tap to the 16 ohm tap of the speaker selector (it has a higher output voltage). That made a significant difference ! The precence control now has a noticeable function and the amp really sings when distortion is up (esp when the precense control is turned down).

Perhaps something that others may want to try ?

Me again:
I found the effect on the tone and amp response from the change I made to the negative feedback so interesting that I now installed a 7 position rotary switch on the back of the amp with different resistors that can go into the nfb loop. The various settings change the character of the amp significantly.

I did the following:
Moved the nfb from the 4ohm to the 16ohm speaker tap. Added a 4.7k resistor in parallel with the 4.7k resistor that is in the original nfb circuit. These two changes should create 4 times more nfb which will be the max setting with the rotary switch.
Then I added the rotary switch with the following values in series of the nfb circuit: 0ohm, 1k, 2.3k, 4.7k, 8k, 20k, open.

So I can now select between much more nfb than original, to no nfb at all in 7 steps.

I first thought a 20k resistor in series would more or less act as no nfb at all, but that is not the case. You can definetely hear the diffrence on the amp behaviour and tone when you switch between no nfb and the 20k, repective to the 8k.

The presence control is still active to change the behaviour of the nfb for higher frequencies (precense appears to be tuned to influence frequencies above aprox 400Hz).

This was not really rocket science but it gives an interesting change to the amp.

One of my favorite mods to any amp with NFB is to add a push-pull switched pot with a series resistor in place of the NFB resistor. Infinite selection in its range and 100% defeatable as well. I usually just put this in by default in any amp I build from scratch.

Look at the AX84 October for a great example of how it's done (that's where I got the idea.)

It's something that's been on my "I'll get to it someday" list for my ODS. Thanks for being the guinea pig

Now that I played around with my new nfb switch in the amp for a few days I realise how fantastic the amp also sounds with no nfb at all when you use no or little overdrive ! The amount of nfb to select really depends on what you want to use the amp for. More nfb is great for some tones, while less is better for some other tones.

Hi @dtp
The reason I did not use a potentiometer for my experiment is that in the OSD the feedback resistor is only 4.7k in series with a 390 ohm cathod resistor to ground of the PI.
This means if the amp is played at its max capacity around 30V could be over the 16ohm speaker terminal and that 4.7k resistor will need to cope with aprox. 200mW... that is not possible for normal potentiometers.. That is probably why the 4.7k resistor is rated 3W in the schematic to be on the safe side. I made the 7 step rotary switch with 2W wishay resistors.

Your garden-variety linear-taper Alpha pots are rated for 1/2-watt... that will easily handle your 200mW current. 3W (and 2W) resistors are overkill here.

Check out this schematic... it's for a 50W power amp, very similar to the OSD power section...
https://ax84.com/archive/ax84.com/stati ... 090501.pdf

Works like a charm.

But your solution works too so it's all good. Just don't be afraid to use a pot next time.

Hi dtp,
Thank you for the link to the interesting amp schematic !
I see a 22k resistor in series with the nfb 100k pot. The 22k resistor limits the max current in that nfb circuit to max 1,3mA if the pot is set to zero and the amp is at max volume (30V at 50W 16ohm).

This was my thinking:
In the OSD the situation is different. The NFB resistor is 4.7k. If we ignore the 390 ohm in series with it (as the 390 ohm is partly paralelled to ground with the cap of the presence dial), the max current in the nfb circuit is 6,3mA at max volume.
I wanted to test to vary the nfb resistor with up to to 20k, so I should then have used a 20k pot.

If a pot can handle 1/2 watt this is the power handling capacity over the whole potentiometer, so when set at its max resistance. If I set a 20k pot at e.g. a 1k setting it can no longer manage 1/2 watt... it can only handle 1/20 of its max power rating as the whole resistance track is not used to dissipate the heat. It is the max current capability of the pot (regardless of its setting) that shold not be exceeded. For a 20k 1/2W pot this should be 5mA. So regardless of the pot setting, as long as 5mA is not exceeded the pot should be fine.

6.3mA of the OSD max nfb current is borderline for such a pot.. In addition I wanted to increase the nfb range while I made this change so I decreased the original 4.7k series resistor of the nfb circuit to 2.3k, which of course doubles the max current...

This is why I concluded that a normal pot will not be suitable here and I better use a rotary switch with fixed resistors. 3W nfb resistor as in the original design and the 2W that I choose are overkill. 1W should be enough at least for the resistors with the lowest value that I connected in series (at max 30V, a 2.3k resistor will burn 390mW).

For the schematic you sent a 1/2W pot looks good as the nfb current appears to be 5 times less compared to the nfb circuit of the OSD. (A 100k 1/2W pot can manage max 2mA. The max current in that circuit should be 1.3mA at 30V with the 22k in series).

Then I was not even sure if the pots I could buy are 1/2 watt rated.. many I found available were only rated 0.125W...

But finally this is perhaps all nonsense as the real output of the amp will not be at 30V max level during long continous times so there is a lot of safety margin...

Also look at the tap the NFB is coming off of... 4 ohm on the OSD, 8 ohm on the October. All of that factors in too. Rob Robinette's website has some interesting info on switching NFB levels in a typical Bassman-type poweramp circuit (which the original D-style amp and the October is derived from.)

https://robrobinette.com/5F6A_Modificat ... ack_Switch

As you can see from that article, with those small NFB resistor values, the OSD has a LOT of NFB which is one of the major contributing factors to its tone and "tactile response" IMO.

In addition to where the nfb is tapped from the speaker selector (highest voltage at the 16ohm tap), I think a bigger difference between these amps is how the nfb is connected at the PI at the cathode.
In the OSD it is connected at a 390 ohm resistor to ground.
In the bassman and the schematic in your previous post it is a 5k resistor to ground.

These reistors work as as voltage divider together with the nfb serial resistor. This is why the OSD has a nfb serial resistor of only 4.7k, while the others have up to 100k..
I guess the relation of this voltage divider of the nfb determines if the nfb needs to be connected to the 4, 8 or 16ohm speaker tab.

Don't forget that your Presence setting has a big affect on the amount of NFB. It all works together.

Using higher value resistors in your NFB/Presence circuit lowers the current requirements of the resistors as well. But even with the small values as used in the OSD, even if you are dropping all 30V (which you are not) the current is only about 10mA.

Like I said, what you are doing works GREAT but, all this to try and convince you... in the future you can use a standard Alpha/CTS/whatever pot with confidence.

By moving the negative feedback signal from the 4 ohm tap to the 16 ohm tap of the speaker selector, you've increased the output voltage in the negative feedback path. This change has resulted in a noticeable difference in the functionality of the presence control and improved the overall sound of the amp, especially when distortion is applied. It's always exciting to discover tweaks and adjustments that can enhance the performance of an amplifier.