hacking kit JTM45 bias circuit

[mikepukmel]

Id like to put the ability in to adjust bias individually on each output tube. Not sure Im on the right track. I think I need to increase the value of the 56K resistor so that the voltage into the pots is higher (larger negative), then feed that value into separate adjustable voltage dividers for each tube. I poked around on the net, didn't see a ready made solution. They all seem to have one pot for both tubes.



Id like to modify values so that I can use off the shelf 10k bias pots (recessed screw) can't find other values.

The hard part is not knowing the current, so I can't use voltage divider and ohms law calculations.

[Enzo]

The current? The only current flowing is through the voltage divider itself.

What voltage is on the first 8uf cap, right at the diode anode? That is across the 15k/56k pair. Ohm's Law will tell you the current there.

You then calculate the voltage at the joint of 15k/56k, the second 8uf cap.. V x 15/56 = ? That is what your pot strings use, so I call that XXX. XXX is across each RB/PB pair. Add up their resistance, RB+PB. Current equals XXX/(RB+PB)

If I start with -60v at the diode, then I have what -47v as XXX. If I assume PB is 10k and RB is oh maybe 20k, then I have 30k. 30k with 47v across means less than 2ma.

Now I cheated there because your RB+PB strings are parallel the 56k. SO I really needed to calculate the total resistance of those two branches and the 56k. But I don't have your values, so I fudged. But the point is that the whole thing draws very little current, so if you are worried if a 2 watt pot is enough or something, most any pot will do.


In a practical circuit, I would surely use larger caps than 8uf. 8uf is a value from a long ago era when caps were expensive, so they used the smallest ones they could get away with. 22uf, 47uf, 100uf sounds better to me.

They all seem to have one pot for both tubes.

I am sure I will get disagreement here, but individual tube adjustments is a fad in amp builder land. The only amps with that are generally high power amps like SVTs that want to keep things under control. Later era amps like some MArshalls have individual controls for the two sides. I think largely because the consumers want them. Like they now put standby switches on things like Champs. The amps only need one control.

Just my personal opinion, but if you want individual bias control, why not add same for screen voltages? And maybe a balancer for the phase inverter? WHen your mains voltage moves a volt, your B+ moves several volts. If your mains moves 5 volts, then your B+ moves 15-20 volts. Your bias will also change too, but the currents won't track for dissipation. SO what happens to the careful bias setup? When we adjust each side, will we compensate for the different resistances in the OT primary halves?

[bob p]

In a practical circuit, I would surely use larger caps than 8uf. 8uf is a value from a long ago era when caps were expensive, so they used the smallest ones they could get away with. 22uf, 47uf, 100uf sounds better to me.

Just my two cents: put in 100uF.

I am sure I will get disagreement here, but individual tube adjustments is a fad in amp builder land. The only amps with that are generally high power amps like SVTs that want to keep things under control.

I think individual tube adjustment becomes important when you're running parallel tubes, especially when dealing with customers who are too cheap to replace a matched sextet. If they blow one tube, they're likely to stick one tube in, whatever they can get their hands on, regardless of how well it matches others in the amp. (Yes, I've dealt with amps that have 6 different brands of 6L6 in them at the same time. ) The problem with that, of course, is that if that one tube isn't well matched with it's parallel partners, one (or two) tubes on that side can end up getting the lion's share of the current while they're doing the heavy lifting, which is bad for tube life and only taunts the next failure to occur. At least with individual bias you can avoid some of the trouble. but I agree -- in a single pair PP amp single bias is nice to have, though not mandatory. IME it tends to be pursued by those who have read that 70% is a magic number and want to tune their amp so that each tube is at exactly 70% at idle. The truth is that biasing a little colder with a fixed resistor and a balance pot setup works pretty well.

[trobbins]

Bias voltage RC time constant may become a concern for hot turn-on if raising the bias cap values. Apart from transformer in-rush and B+ filter cap in-rush, the output stage would also be pretty much fully loading the B+ at that time.

Some owners may not have many output stage tubes, and could end up needing to use tubes with quite different bias current levels (which they would have to use a bias probe to check, as there is no cathode sensing capability by default). I'd certainly be happier if I could remove large disparity if I new that was the situation.

[Pedro Vecino]

This is the system Marshall uses in the 1992 bass model (the modern one with slope, sweep, etc. within the JCM800 series):

[Enzo]

There, much better, and less parallel crap to figure.

[mozz]

Bias voltage RC time constant may become a concern for hot turn-on if raising the bias cap values. Apart from transformer in-rush and B+ filter cap in-rush, the output stage would also be pretty much fully loading the B+ at that time.

Some owners may not have many output stage tubes, and could end up needing to use tubes with quite different bias current levels (which they would have to use a bias probe to check, as there is no cathode sensing capability by default). I'd certainly be happier if I could remove large disparity if I new that was the situation.

In other words, make your bias cap small than your first filter cap. In this instance, the first cap is 32uf. You want your bias cap at 22uf. From a cold start you want you bias voltage there before the HV on the plate gets there. Here are a few Hiwatt bias circuit mods.

[mikepukmel]

The current? The only current flowing is through the voltage divider itself.

What voltage is on the first 8uf cap, right at the diode anode? That is across the 15k/56k pair. Ohm's Law will tell you the current there.

You then calculate the voltage at the joint of 15k/56k, the second 8uf cap.. V x 15/56 = ? That is what your pot strings use, so I call that XXX. XXX is across each RB/PB pair. Add up their resistance, RB+PB. Current equals XXX/(RB+PB)

If I start with -60v at the diode, then I have what -47v as XXX. If I assume PB is 10k and RB is oh maybe 20k, then I have 30k. 30k with 47v across means less than 2ma.

Now I cheated there because your RB+PB strings are parallel the 56k. SO I really needed to calculate the total resistance of those two branches and the 56k. But I don't have your values, so I fudged. But the point is that the whole thing draws very little current, so if you are worried if a 2 watt pot is enough or something, most any pot will do.


In a practical circuit, I would surely use larger caps than 8uf. 8uf is a value from a long ago era when caps were expensive, so they used the smallest ones they could get away with. 22uf, 47uf, 100uf sounds better to me.



I am sure I will get disagreement here, but individual tube adjustments is a fad in amp builder land. The only amps with that are generally high power amps like SVTs that want to keep things under control. Later era amps like some MArshalls have individual controls for the two sides. I think largely because the consumers want them. Like they now put standby switches on things like Champs. The amps only need one control.

Just my personal opinion, but if you want individual bias control, why not add same for screen voltages? And maybe a balancer for the phase inverter? WHen your mains voltage moves a volt, your B+ moves several volts. If your mains moves 5 volts, then your B+ moves 15-20 volts. Your bias will also change too, but the currents won't track for dissipation. SO what happens to the careful bias setup? When we adjust each side, will we compensate for the different resistances in the OT primary halves?

Enzo, Re "why not add same for screen voltages? And maybe a balancer for the phase inverter? WHen your mains voltage moves a volt, your B+ moves several volts. If your mains moves 5 volts, then your B+ moves 15-20 volts. Your bias will also change too, but the currents won't track for dissipation. SO what happens to the careful bias setup? When we adjust each side, will we compensate for the different resistances in the OT primary halves?"

Oh man, don't get me started! Im a tinkerer!! Thanks, you just fed the fire! Every resistor should have a pot or trimmer in series.

Seriously, thanks for the tech info. I dont have the amp built yet, so I don't have any voltages to measure. Just thought I should put the parts on the mouser order list Im building up. I bought a JTM45 kit, got here a couple of weeks ago. The transformers should be here end of the week. Now that I have had time to sit and brood and stew over grounding, which brand of tubes, etc, bias hit me.

The kit comes with a kind of cheesy 22k trim pot. I looked around on the web, could not find anything that looks like the ones used in Fender amps, so i started thinking maybe its time to think (aka brood, and stew) over not being able to adjust each tube's bias individually. On my Fender deluxe type amp, it runs fine, but the output tubes even though they were purchased as a matched pair, run 'different enough'. one glows a little brigher than the other, and the currents are not with in 0.0001% like I like them to be <sarcasm off>. Joking aside, the output transformer dc resistance measurement for the two halves of the OT secondary were different by more than I expected. I think that caused part of the bias difference. Not only the tubes, but the half of the OT winding that I have no control over. (or maybe not, and all this is for naught).

[mikepukmel]

In other words, make your bias cap small than your first filter cap. In this instance, the first cap is 32uf. You want your bias cap at 22uf. From a cold start you want you bias voltage there before the HV on the plate gets there. Here are a few Hiwatt bias circuit mods. [ATTACH=CONFIG]47406[/ATTACH]

Cool,. thanks for the tips and schematic clips. Some year fender's had a 70uf (I think) bias filter cap, but still had much smaller filter caps. I can't remember now, have to look. For cold start, you mean, without letting the amp warm the heaters up via the standby switch, i.e. accidentally leaving the standby off, and turning the amp power on? eek.

[mikepukmel]

This is the system Marshall uses in the 1992 bass model (the modern one with slope, sweep, etc. within the JCM800 series):

[ATTACH=CONFIG]47405[/ATTACH]

Thanks Pedro!

[mikepukmel]

I just checked the kit parts: the first two filter caps (before and after the choke) are 32uf. The bias caps that the kit came with are 10uf. I have some 25uf/50v but probably not enough voltage here.

Not sure about tolerances, these are I think 20%. So, max has to be below 32uf * (1 - 0.2) = 26 so I can go with around 22uf/100v?

[bob p]

Telling us that the caps are 32uF isn't enough. We need to know if they're totem-poled or not. If they are, it cuts their value in half. Unfortunately I don't see them on your schematic clip.

Tiny bias caps went away in the 60s when caps became cheap. FWIW a great many Fender amps commonly used bias supply caps that were larger than the main PSU caps; AB763 BF Twin Reverb (50uF vs 35uF); Super Twins: (80 uF vs 50uF); AA371 Bassman (50 vs 35). My Super Twin got "upgraded" to have 100uF on the bias supply with 50uF on the first filter pole. It hasn't gone up in flames ... yet.

My experience with modeling bias and PSU supply activity shows that the caps come up to voltage pretty darned fast -- typically within an insignificant number of cycles. I think you'd have to have some crazy long time constants before you'd see tubes incinerating at power on. Has anyone actually seen this happen?

edit: Those HiWatt circuits in Mozz's diagram used some pretty large resistor values (82k, 110k) with 100uF bias caps.
What size caps are on the PSU nodes?

[Enzo]

The last thing you need to worry about is tolerances. Remember the stock circuit in your schematic uses 8uf. You already found the OT is not "balanced". 8uf would work, bob's 100uf would work, and anything in between would work. It just isn't critical.

I have slightly different thoughts on relative size of cap. The high voltage caps tend to be smaller, and they charge off the rectifier. Your bias cap only has to charge up to like 50v to be enough to move on, but it is doing that through a 180k resistor. I would have to think that resistor rather than the relative cap size is what slows it down. But then one has to wonder how many seconds does it take the bias to charge up to tube protecting levels, compared to the 10-20 seconds it takes the tube heaters to allow current to flow.

My suggestion to make EVERYTHING adjustable was tongue in cheek. I was using some friendly sarcasm to make the point that making ONE part of the formula precision doesn't change that the whole thing is not remotely precision. It certainly won't hurt the amp to make all the bias adjustments independent. I just think it is wasted energy. But that is only my opinion.

[J M Fahey]

In other words, make your bias cap small than your first filter cap. In this instance, the first cap is 32uf. You want your bias cap at 22uf.

Sorry but it doesnīt work that way.
Cap value per se is irrelevant, what matters is an RC time constant and you are ignoring the R part of the equation.
Mains filter caps will charge in one or two seconds, because transformer has a very low internal resistance, think 10 to 30 ohms
Just suppose worst case 100uF mains caps and 50 ohm winding resistance, time constant is 0.005 seconds (not a typo, 5 milli seconds).
While best case 180k and 8uF means 1.44 seconds.
I use and recommend:
RC pad corner frequency upper and lower cutoff frequency calculation filter calculate time constant tau RC voltage power calculator capacitance resistance - sengpielaudio Sengpiel Berlin
Even worst case 180k and 100uF would mean very acceptable 18 seconds.
Just donīt go overboard and use a Trillion Farad Flux Capacitor there , put it back into your custom made De Lorean

[pdf64]

...the output tubes even though they were purchased as a matched pair, run 'different enough'. one glows a little brigher than the other, and the currents are not with in 0.0001% like I like them to be <sarcasm off>. Joking aside, the output transformer dc resistance measurement for the two halves of the OT secondary were different by more than I expected. I think that caused part of the bias difference. Not only the tubes, but the half of the OT winding that I have no control over. (or maybe not, and all this is for naught).

How brightly a tube glows is not a specified characteristic, it's irrelevant for the application.
Unless an OT is manufactured using an expensive 'bifiliar' technique, given an equal number of turns, the legs of each primary side will tend to have different resistances, as the inner turns will require less wire than outer turns.
And with regard to a power tube's operating conditions / plate current, the plate voltage is far less significant than its screen grid voltage.