Bias Question

Collings

I want to mod an amp that has no bias test points so I can check the bias easily.

I'm going to add a couple of one ohm resistors between the valves cathode and ground but what I need some advice on is what wattage these need to be?

I've brought 1 ohm 1% 2W - are they going to be ok?

Thanks

ICBM

Yes, perfect.

1W is the minimum since a shorting valve can pass up to about 1A before the amp's fuse blows, which produces 1W in a 1-ohm resistor obviously - so 2W is a good safety margin.

ecc83

Whilst it is nice to have a resistor per OP valve there is not a lot of point unless there is provision to balance the cathode currents with dual bias pots.  In any case such niceties are a bit wasted on guitar amps!  As a rule the cathodes are commoned down one resistor, this also make the maths easier! One volt = one amp as IC says and so of course mV =mA.

If you want belts AND braces, stick a 1A diode across the resistor.

Dave.

Collings

Thanks both.

jpfamps

I would always put a resistor on each cathode so you can monitor the current through each valve.

That way if your valves aren't well matched then you can make sure that they aren't over dissipating as you can calculate 

Using a common resistor for all the valves then you don't know what each valve is dissipating and say if one valve is not conducting at all you may end up biasing the working valve/s very hot.

I've seen this in several amps, including one version of the SVT.

RiftAmps

I always use a resistor for each valve, but only because they cost a few pence each and I like to get down 'n' nerdy with my sums. If you can trust your valve supplier to deliver actual matched valves (both PC and TC) then I suppose a common R will suffice.

Here's a common way to do it with external bias points (note sockets wired for EL34s).

ecc83

RiftAmps said:

I always use a resistor for each valve, but only because they cost a few pence each and I like to get down 'n' nerdy with my sums. If you can trust your valve supplier to deliver actual matched valves (both PC and TC) then I suppose a common R will suffice.

Here's a common way to do it with external bias points (note sockets wired for EL34s).

Yes, I agree, nice to do and the 'right way' to do it!

However, the average Joe is not going to have a rake of OP valves to try so even if the valves are a bit mismatched there is little he can do about it (could fit a bias balance pot* of course) . Ignorance is bliss?

The consequences of mismatched valves will be an increase in remnant hum and (totally inaudible for guitar!) a bit more even harmonic distortion.  Of course , a SERIOUSLY out of specc' valve will hum badly.

Large production runs will use batches of valves which 'should' be fairly consistent and a bad imbalance should fail a S/N test.

There is a bit of a paradox here. Cathode biased guitar amps almost never have individual Rs and caps whilst Hi Fi amps almost always do. Most valve data sheets also require separate cathode resistors. If we wanted to have equal currents in the valves, 'we' would fit two resistors but nobody, AFAIK ever does!

*Bs of course fit a balance pot on most of their range. Never seen a real need?  The OP stage balance is set with a special meter that reads low frequencies but Mk 1 lug close to the speaker works just as well!

Dave.

ICBM

There are a very few guitar amps which use separate cathode bias - the Vox AC50, and at least one of Dennis Cornell's models for example - but it isn't common.

One reason it may not be is that it doesn't 'sound right' if you want a typical cathode-bias sound. It sounds more like fixed-bias in some ways, which is apparently why Cornell used it on his mini Plexi.

I first came across this a long time ago when a friend had an AC30 which a hi-fi tech had 'improved' by fitting separate cathode resistors, caps and trimmers for each valve, so they could all be biased perfectly. It sounded somehow flat and lifeless, and not like an AC30 - so as a quick test I replaced the beautifully-made little circuit board with a single resistor and cap... instant great AC30 tone.

Separate bias-metering resistors are the best method though, as jpfamps said - with a single one you don't know if one valve is drawing too much (or all) the current.

ecc83

"Separate bias-metering resistors are the best method though, as jpfamps said - with a single one you don't know if one valve is drawing too much (or all) the current. "

Well, you would IC coz it would hum like a cnut! N very B I am not saying that separate sense Rs are a bad thing at all, just a bit OTT for guitar amps.

Dave.

ICBM

ecc83 said:

"Separate bias-metering resistors are the best method though, as jpfamps said - with a single one you don't know if one valve is drawing too much (or all) the current. "

Well, you would IC coz it would hum like a cnut! N very B I am not saying that separate sense Rs are a bad thing at all, just a bit OTT for guitar amps.

I still disagree, from practical experience.

Suppose you have a four-valve amp, with a common bias sense resistor for both valves on each side and a bias trimmer for each side. If one valve is not working you will end up biasing the other one on the same side to double the correct current, and there will be no hum.

This is not a theoretical example, it's a Marshall DSL/TSL100…

OK, with a two-valve amp or a four-valve one with a single resistor then it would still hum.

But separate resistors are always better and there is no downside, other than very marginal cost - they will also allow you to easily spot a badly-matched valve set, which is quite common even if they're supposedly matched.

ecc83

I bow to your practical experience IC but, "Bias trimmer per side" How many 4xEL34 (say) 100W amps have those?

Of course, IF you have the ability to balance the bias voltages then yes, of course you must have individual sense resistors but I i say again, how many  (old) PA and guitar amps had the facility to balance the valves?

I agree that the cost of the extra sense Rs is minimal (but NOT trivial! see in a minute) but then you really have to have two bias pots (or do it 'our' way) . I have said this many, many times in Frets, Guitar amps is a VERY competitive game! I would guess adding another 2W 1R plus preset (and don't forget the PCB has to be designed for it!) would add about $1 per chassis and a $ in production means $10 in the shop and THAT can put you into another price range, i.e. the wrong side of a grand!

In any event, in these days of super accurate DMMs you can get a very good idea of relative cathode current by careful inspection of Vg2 a technique I used to determine the 'odd' valve in A30s when they seemed a bit hummy.

I WILL admit to being a bit anal about amp hum! A relic I guess of my Hi Fi history.

Dave.

ecc83

BTW. I will agree about 'supposed to be matched'. I went through about ten sets of pairs and quartets of EL84s and EL34s and found many were further out of their 'group' than others that were not in it (iyswim!)

I see no solution to this problem for the cash strapped player? Back in the day, this is the sort of thing the magazines would take on? Buying in a selection of so called 'matched sets' and then getting them professionally tested.

Nowadays mags do not have the money, staff or time and are scared witless of losing advertising.

Dave.

ICBM

ecc83 said:

Of course, IF you have the ability to balance the bias voltages then yes, of course you must have individual sense resistors but I i say again, how many  (old) PA and guitar amps had the facility to balance the valves?

Not many. Actually, I prefer to have a single trimmer and separate resistors on the grounds that if the valves are properly - or at least reasonably - matched, then there is no need for separate bias adjustment. It's not a good idea to force a balance by having completely independent bias with very mismatched valves in my opinion - since if the idle current is that different, it's likely that the transconductance will be as well. Of course that seemed to matter little in the old days too, when many amps didn't even have bias adjustment at all! But then again, valves were usually more consistent too…

The reason Fender moved to bias-balance rather than bias-adjustment in the late 60s is because they could then buy ungraded valves in bulk knowing that they were all close to the design spec so the overall bias would be about right, and get a perfect hum-free balance with any set. Nowadays bias adjustment is more useful because the valves vary so wildly.

clarkefan

ICBM said:

ecc83 said:

"Separate bias-metering resistors are the best method though, as jpfamps said - with a single one you don't know if one valve is drawing too much (or all) the current. "

Well, you would IC coz it would hum like a cnut! N very B I am not saying that separate sense Rs are a bad thing at all, just a bit OTT for guitar amps.

I still disagree, from practical experience.

Suppose you have a four-valve amp, with a common bias sense resistor for both valves on each side and a bias trimmer for each side. If one valve is not working you will end up biasing the other one on the same side to double the correct current, and there will be no hum.

This is not a theoretical example, it's a Marshall DSL/TSL

@ICBM, wouldn't this mean then that the DSL *50* is the best setup? Separate adjustment for each of the 2 power valves?  I thought the DSL biasing points thing was frowned on, maybe it doesn't have separate sense resistors, I'm not up on this stuff...

jpfamps

I concur with ICBM.

I like to have a single bias adjust, but have the ability to measure the cathode current through each valve.

That way you can be sure that none of your valves are over dissipating and all are drawing current.

If your valves are so badly matched that they can't be biased satisfactorily with a single adjustment then they probably shouldn't be used.

A common problem when valves short is that the screen grid resistors go open circuit, so any new valves installed in that position will not conduct. In a 4 (or even 6) valve output stage with single current sense resistors you would end up biasing a remaining valves too hot.

I have actually seen this in several amps, including some versions of the SVT and T/TSL100s where the user has changed that valves and followed the biasing procedure in the manual.

I've also seen this in HRDs with a single bias test point. You would be surprised how little hum there is in these amps with only one valve conducting.

jpfamps

clarkefan said:

ICBM said:

ecc83 said:

"Separate bias-metering resistors are the best method though, as jpfamps said - with a single one you don't know if one valve is drawing too much (or all) the current. "

Well, you would IC coz it would hum like a cnut! N very B I am not saying that separate sense Rs are a bad thing at all, just a bit OTT for guitar amps.

I still disagree, from practical experience.

Suppose you have a four-valve amp, with a common bias sense resistor for both valves on each side and a bias trimmer for each side. If one valve is not working you will end up biasing the other one on the same side to double the correct current, and there will be no hum.

This is not a theoretical example, it's a Marshall DSL/TSL

@ICBM, wouldn't this mean then that the DSL *50* is the best setup? Separate adjustment for each of the 2 power valves?  I thought the DSL biasing points thing was frowned on, maybe it doesn't have separate sense resistors, I'm not up on this stuff...

The DSL biasing system is not ideal because adjusting the bias for one valve affects the bias on the other side!

I would have preferred a single adjustment and 2 test points.

Having said that the bias supply not the worst problem with the DSL.....

ecc83

I can tell when I am in a hole so I shall stop digging!

I guess the reason I mentioned increased hum was because I had to check it against a specification. I have no idea what the S/N spec would be for the other makes mentioned!

Maybe older guitar amps did not have bias pots but decent valve PA amps did. In fact the Wimbledon (2X KT66) had a METER on the front that allowed you to measure and check Vss, Va and anode current. Oddly perhaps THE most famous KT66 amplifier, the Quad ll had but a single cathode resistor, but I am bloody sure Peter Walker knew what he was doing!

Dave.

ecc83

"The DSL biasing system is not ideal because adjusting the bias for one valve affects the bias on the other side!"

Oh! I wonder why they did not use the earthed wiper system where you get a differential bias adjustment? There is a little bit of interaction but not a lot.

Dave.

ICBM

The DSL/TSL bias scheme is also extremely poor design because unlike the older 'pull-down adjust' method in vintage Marshalls where a trimmer failure causes the voltage to go *up* - ie fail safe - in the DSL/TSL the bias supply passes through both a jumper connector (twice) and the bias trimmer, so a poor contact at any of these three points results in no bias voltage and blown power valves. This is the result of a jumper connector fault in a DSL100…



Just one of the many design faults in these amps by someone who clearly didn't know much about valve amp failures.