Help improve pp-18 tone

18watt-specific Tech Talk - Building, Fixing, Parts, Mods...

Moderators: CurtissRobin, colossal, zaphod_phil, Daviedawg, Graydon

TomOlsen1
Occasional poster
Occasional poster
Posts: 159
Joined: Thu 09/18/08 2:00 am
Location: Jo´burg, South Africa

Post by TomOlsen1 »

Would it not be possible to use a variation of THIS to make the zener track the VVR? I'm not conversant enough in electronics to see if it'll work.

http://www.electronicecircuits.com/elec ... uper-zener


-Ewald
0 x

User avatar
zaphod_phil
Builder, Admin
Builder, Admin
Posts: 15208
Joined: Wed 03/19/03 2:00 am
Location: YYZ

Post by zaphod_phil »

Would it not be possible to use a variation of THIS to make the zener track the VVR? I'm not conversant enough in electronics to see if it'll work.

http://www.electronicecircuits.com/elec ... uper-zener -Ewald
No, I'm afraid not. All that does is provide a reference voltage, which isn't what the zeners are being used for here. I think your multi-way switch is th best alternative I've seen so far. The leads to the switch need to be kept real short so that they don't radiate PA drive signal into the rest of the amp.
0 x
Nature abhors a clean tube amp

User avatar
asd
Newbie
Newbie
Posts: 75
Joined: Fri 08/28/09 2:00 am
Location: Unknown

Post by asd »

0 x

katopan
Frequent poster
Frequent poster
Posts: 914
Joined: Mon 12/10/07 2:00 am
Location: Melb, Aust
Contact:

Post by katopan »

A linear regulator, not what we're after either. I don't believe a variable Ruby zener mod has been done before, but you need a variable threshold hard clipping circuit that is able to sink current. Have played with such things on video signal processing circuits using a transistor around the feedback path of an opamp. When below the threshold it doesn't draw current, above the threshold it draws as much as the impedance of the source allows to flow to keep the voltage clipped. You do get transition spikes that could be a problem, but I'm wondering if a simpler version would be possible using a pot and a transistor. Damn that attachments don't work anymore... picture tells a thousand words.
0 x

TomOlsen1
Occasional poster
Occasional poster
Posts: 159
Joined: Thu 09/18/08 2:00 am
Location: Jo´burg, South Africa

Post by TomOlsen1 »

As luck would have it, I'd drilled the hole for the rotary as close as I could to the PI, and the wires are as short as possible and sitting well away from the PI and PA sections. I've even got them twisted. I'll try post some pics later. My 18 is extremely quiet after my recent rebuild. So much so that I'm considering giving my Plexi 6V6 the same kind of treatment.

Hehe, I think this is the first thread where Ewald and myself are both posting from my ID.

Regards,
Tom
0 x

User avatar
asd
Newbie
Newbie
Posts: 75
Joined: Fri 08/28/09 2:00 am
Location: Unknown

Post by asd »

Unfortunately this is way beyond my current electronics knowledge.
If it was software it would be super easy : )

In any case, I'm very satisfied with my tone ATM (basically 220k grid leaks and warmer PI), and I'm wating for the NOS tubes to arrive.

However, it would be great to have something like that.

What about this circuit? http://www.daycounter.com/Circuits/Vari ... ener.phtml

As I understand it, as R1 and R2 value changes, the threshold of the "zener circuit" changes too.

If this is the case, it would just be a matter of using a dual ganged pot in the VVR (with correct values) and use it to change the threshold.

Sure enough I don't have a clue about how to simulate this thing or which transistor to choose...
0 x

katopan
Frequent poster
Frequent poster
Posts: 914
Joined: Mon 12/10/07 2:00 am
Location: Melb, Aust
Contact:

Post by katopan »

Yes, that's the sort of thing I was thinking of. If you set the resistor ratio to track right with the pot movement you could connect the V+ to the Vout so you don't even need a separate supply and have it switch in and out of conduction at the right (variably tracked) threshold. Bit crude, and yes it'll drift a bit with temperature, yada, yada. But could possibly work and is very simple.

The Ruby zener clamping comes into play on the negative side driving into the power grid, when that side power valve is cut off. Any conduction spikes from the variable zener wouldn't be directly audible but you'd need to be careful that it doesn't induce conduction spikes into nearby signal paths.
0 x

User avatar
asd
Newbie
Newbie
Posts: 75
Joined: Fri 08/28/09 2:00 am
Location: Unknown

Post by asd »

So if I tie V+ and Vout together, this node will be connected to R2, C2 and R5 (http://www.18watt.com/storage/18-watter ... fo_311.pdf) and the ground to the diode ?
Now I need 2 of them, and I need to track the right voltage. I should use SPICE or something?
0 x

katopan
Frequent poster
Frequent poster
Posts: 914
Joined: Mon 12/10/07 2:00 am
Location: Melb, Aust
Contact:

Post by katopan »

OK, looking at this in a bit more detail I said a couple of wrong things before. Firstly the zeners in the Ruby mod conduct and clip the negative side of the grid signal and so source current rather than sink it. To make the variable zener circuit source current you need to change the PNP transistor over to an NPN.

Secondly I said that you could tie V+ & Vout together, but that would get rid of the control voltage (V+) that sets the conduction (effective zener) threshold. But the problem here is that setting a negative clipping threshold requires a negative V+ voltage, something we don't have in an 18 Watt. Or if we do tie them together then the variable threshold would need to be set by changing the R1/R2 ratio with a pot. Ideally it'd be great to have this thing track properly based around having a dual gang VVR pot and using the second half to track this variable zener circuit.

Needs a bit more thinking through but I think you're on to something asd. Very much uncharted territory too! Up until now it's just been accepted that the Ruby mod doesn't work as you turn down the VVR. Simulation is probably not necessary. Power valve bias tracks pretty well with VVR output and the Ruby mod isn't terribly sensitive to the threshold voltage so tracking with VVR off a dual gang pot would probably work OK and be perfect in ease of use. Just have to be careful with layout and not running the power grid signals all over the place. Or at worst a VVR pot and Ruby mod pot which you have to set to the same position.

Forget if you mentioned it, but what sort of VVR are you running? (whole amp, PA only or PI/PA) And I think you mentioned before that with the VVR up full the Ruby mod does fix your amp buzz?
0 x

User avatar
asd
Newbie
Newbie
Posts: 75
Joined: Fri 08/28/09 2:00 am
Location: Unknown

Post by asd »

First of all thanks for your interest. And yes I'm running PI+PA VVR.
I run the Ruby fix once but Ihadn't had the chance to play it loud (didn't understand back then that it doesn't work with the amp scaled).
In any case I assume it should work since I tried any other buzz fixing mod, and I didn't get rid of the buzz.

In any case are you saying that I just need to use a NPN transistor, tie the V+ and Vout together, then use a pot instead of a resistor?

And if I tie V+ and Vout together, will be Vz= (R2*V)/(R1+R2)+Vbe still valid?
0 x

katopan
Frequent poster
Frequent poster
Posts: 914
Joined: Mon 12/10/07 2:00 am
Location: Melb, Aust
Contact:

Post by katopan »

It would be nice to know if the Ruby mod gets rid of the buzz in your amp with the VVR up full. Actually if I read it right looking back, the fizz comes in as you wind the VVR down? Anyway, happy to proceed but just keep in mind that if you're not getting the buzz that the Ruby mod fixes with the VVR up full, then the variable zener thing may not do what you are after either.

But moving on.... firstly I need you to do a test. Please measure your B+ and the power cathode voltage with the VVR both up full and down at minimum. I know what these values are for my amp, but it'd be good to have them for your amp for base the resistor/pot design around. Also are you using a linear (more common) or log pot for VVR control?
0 x

Daviedawg
Superior Amp Tech
Superior Amp Tech
Posts: 748
Joined: Fri 01/08/10 2:00 am
Location: Scotland

Post by Daviedawg »

I fitted the Ruby mod after fitting VVR to PI/PA stages. It was intended to reduce fizz at lower voltages. I seldom if ever use the amp at the volumes it was designed for which is a serious shame.
Also I never use it at the minimum voltages available from VVR because it does not sound as it should.
My Normal Hi input is paralled. Using that input generally I have the VVR set at about 25% to 35% of its sweep which is about 80V to 115V on the plate of the EL84s. It sounds like it should with no fizz. But I cannot get a good tone with my tube screamer and humbuckers at that level. That seems to send it over the top into less musical distortion. It varies with the other channels due to lower signal levels presumably. Also at higher voltages using the TS is supremely good for drive and sustain.
I have not at any stage done proper measurement of the outcome. I also note what Phil says above. But I am satisfied that fitting the Ruby fix sorted my fizz issue down to just a few watts of output at a level which is fine for regular home use. Below that other tonal issues come into play anyway.
There are of course lots of variables such as pick ups, guitar volume, etc. This is just my reflections on the topic.

Dd
0 x

User avatar
Gary
Newbie
Newbie
Posts: 93
Joined: Fri 08/18/06 2:00 am
Location: Leeds - as in "Live At ..."

Post by Gary »

zaphod_phil wrote:When you VVR-scale the PI+PA, you're meant to use the amp's Volume control to educe the preamp drive to the PI by a corresponding amount. Don't try to turn up the control to the same level as with the VVR maxed.
I wonder to what extent we vvr devotees are expecting too much from the circuit. Could it be that some of us are causing the fizz by not observing Phil's suggestion. Would the amp/speakers still be fizzy at maximum vvr (gigging) levels - probably not.
0 x

User avatar
asd
Newbie
Newbie
Posts: 75
Joined: Fri 08/28/09 2:00 am
Location: Unknown

Post by asd »

It could be, but I clearly had the buzz when I tried the amp at full voltage (se the other thread of mine), even if I must say that in a band situation, it isn't so annoying, while if you sit with your hear near the speaker it's quite noticeable.

Anyhow katopan I'll reinstall the Ruby fix before wednesday when I have a rehearsal, so we'll see if I'll get the buzz sorted at max voltage.

For the measurements on my amp, unfortunately I have to wait tomorrow (dead battery on my DMM : ) )
0 x

katopan
Frequent poster
Frequent poster
Posts: 914
Joined: Mon 12/10/07 2:00 am
Location: Melb, Aust
Contact:

Post by katopan »

I agree with what everyone is saying. I've got a 68k resistor to ground in my VVR (instead of the normal 100k). B+ goes down to 25V and it sounds different to up full, but suits my needs at home and the distortion at that extra low volume still sounds better than a pedal or SS amp to my ears. I am happy to be "fizz free" up loud or down at TV levels and so actually have never tried out the Ruby fix.

But the idea of a variable Ruby zener mod to track with the VVR has been mentioned and dismissed a few times. It's about time we gave it a shot! :wink:

asd - My insistance in verifying if the Ruby mod fixes your issue with the VVR up full is just because for some people it works like a charm, others have a different fizz issue that the zeners do nothing for and they normally have success with a conjunctive or Zobel filter. Just want to make sure we're heading down the right path for your amp issue rather than wasting your time (and mine). Look forward to hearing back about those measurements.

Tom Olsen has already said the multiple zeners on a switch works for him, so that's a good option too. But the transistor based variable clipping circuit (variable zener) looks like it'd work too if we get the design right.
0 x

User avatar
asd
Newbie
Newbie
Posts: 75
Joined: Fri 08/28/09 2:00 am
Location: Unknown

Post by asd »

Sorry for the delay:

-unscaled: B+=330V, kV=9.85v

-scaled: B+=30V, kV=0.72V

The pot is the stock shipped by Dana Hall, it should be 1M linear pot (it has a circled "a" on it).

I installed a 12v zener Ruby mod, and the fizz seems to go way at max volume, while it's quite evident at medium VVR settings.
It seems that if I turn the VVR down enough (max attenuation) the fizz goes away again. Does this make sense?

And I found that if I strike a chord, kV goes up to 15-16 V. Does this mean that, since the bias shift to the cold side, I'm also getting less power? Could a bigger cathode cap give more output power?
I would be glad if the amp was a little bit louder : )

I also found true what Daviedawg said.
0 x

katopan
Frequent poster
Frequent poster
Posts: 914
Joined: Mon 12/10/07 2:00 am
Location: Melb, Aust
Contact:

Post by katopan »

Variable current source with a negative threshold voltage ranging from 12V down to around 1V, pot controlled. I'll work out the ratios and get back to you. Easy to try it out on another pot. Bit harder to match it to a dual gang pot installed on the VVR.
0 x

User avatar
asd
Newbie
Newbie
Posts: 75
Joined: Fri 08/28/09 2:00 am
Location: Unknown

Post by asd »

That's so kind of you, thanks a lot!!

And... NOS tubes in the mail!! I'm putting them in this evening, I was thinking to increase cathode resistor to dissipate about 10W and to add 100ohms screen resistor as suggested here, and maybe to decrese screen voltage a bit as suggested here: http://www.18watt.com/storage/18wattLite2b.gif

Is there anything else I should know to preserve NOS tubes life without sacrifing the tone too much?
0 x

User avatar
zaphod_phil
Builder, Admin
Builder, Admin
Posts: 15208
Joined: Wed 03/19/03 2:00 am
Location: YYZ

Post by zaphod_phil »

I was thinking to increase cathode resistor to dissipate about 10W and to add 100ohms screen resistor as suggested here, and maybe to decrese screen voltage a bit as suggested here: http://www.18watt.com/storage/18wattLite2b.gif
In that case increase the cathode cap size to 1000uF+, to help regulate the bias voltage.
0 x
Nature abhors a clean tube amp

User avatar
asd
Newbie
Newbie
Posts: 75
Joined: Fri 08/28/09 2:00 am
Location: Unknown

Post by asd »

I'll do!!
0 x

Post Reply