Full-wave rectifier for JTM45 Dual Bias Circuitry

[Steve_Hogan]

This is my first post in this group -- looks like a nice place to share and learn.

I am providing tech support for a friend who is building a JTM-45 Tremolo "Bluesbreaker" amp.
The goal is to match the tone quality as much as humanly possible to the amp(s) used by Clapton, but to eliminate as much hum and buzz and other problems that accompanied the amps of that day. We are using NOS GEC KT66, NOS Mullard 12AX7, NOS mustard caps, etc, etc. Due to the expensive tubes, I intend to employ either dual bias or bias + balance controls on this amp. We plan to bring test points to the rear panel and remote the bias and/or balance pots to the rear panel as well, so the bias can be checked and adjusted without removing the back cover of the case. Maximum reliability to protect the output tubes is desired.

Due to the low cost of 1N4007 rectifiers today, I cannot understand why the bias circuit doesn't use full wave rectification instead of 1/2 wave -- 2 diodes instead of 1.
As I see it, there are many advantages:
1. Significantly lower ripple due to 100/120 Hz charging of the 10uF caps instead of 50/60 Hz with resulting lower hum.
2. Faster charging of the bias circuit when the amp is first turned on which can only help.
3. Failure of a single rectifier (admittedly unlikely), will still result in some Bias to prevent tube destruction.
4. Very slightly less DC in the PT due to balanced "withdrawal" from the bias circuit.

The more frequent charging of the caps will likely result in a change in the voltage, so some values may have to be adjusted.

Can anyone help me understand if there is a down-side to this approach.

[Steve_Hogan]

After submitting my post above, I spent some time with spice to figure out some values. The full wave rectifier does a much better job of charging the first 10uF cap, so the starting negative voltage is higher and comes up quicker (about 2 seconds) than the 1/2 wave rectifier. I would think that the sooner the bias comes up, the better because you want the bias to be fully up to speed when the filaments get hot and the tube begins to conduct.

I started with 2 1N4007 (or even better, SiC schottky low-noise diodes) located on or near the GZ34 tube socket. Both anodes are connected together to form a full wave rectifier with negative output relative to 0V. From that point there are two completely independent and identical bias networks consisting of:
220K series resistor, 10uF to ground, 20K series (was 15K), 2nd 10uF to ground, 33K (was 56K)in series with 20K bias setting pot to ground (was 25K). Series resistor to grid is 220K + 5K6 grid stopper right at pin.

This network provides a range of -38.8 Volts nominal to -58.0 volts nominal. Is this range suitable for a GEC NOS KT66?
I can rework values to put the bias pot into the correct range. If wiper goes open circuit in my circuit, the Bias goes to maximum negative -- lowest Output tube current to protect the tubes. Should one rectifier fail the bias will be less negative, but will not disappear. I personally think this approach is simpler than a bias + balance control to design and build, but is no more difficult to adjust.

I am open to correction here. I've designed, modified and built lots of audio gear for many years, but this is my first JTM-45 Bluesbreaker build.

[dai h.]

hi, I'm not an expert so I don't quite know what the correct answer is, but my understanding is that for the bias circuit you just basically need a voltage and there isn't much current draw (so seemingly no need for stringent ripple reduction as in lots of filtering, more stages, regulation and whatever). And the push-pull configuration maybe helps further by hum cancelling(?). As far as indiv. bias controls I think that is a very good idea in general (and if using old stock tubes especially nowdays with those only getting scarcer as time goes by).

[neikeel]

dai is too modest - he knows more than he is letting on!

You have to be careful when modding these amps. Remember it is far from hi-fi and the reason that they sound so good and give the tones that they do is that they are made of proprietary equipment available in the early 60's at a god price. Design having been borrowed from earlier Fender (and other designs).

The valve rectifier when working in conjunction with the slightly underspecced PT gives a significant sag in the power supply, so do not go for a bigger better modern PT and extra rectification - because more is better - it will not! same goes for the rest of the amp, fill it full of high spec mfs and polystyrene caps and you will have nothing like the organic BB tone you seek.

But yes it will cope with NOS GECs in that range!

[Brian Wallace]

I was always under the impression that the bias circuit was such a low current application that it didn't require anything more than a single diode. Basically, it is not suitable to noise due to it's low current draw. Of course if one of the bias supply filter caps goes bad you will hear the hum come through for sure. If anything would cause some noise in the power section I would suspect of the tube rectifier 5V tap as this drops 5VAC directly onto the power rail that feeds the entire amp. With all this said, I haven't had any real issue with hum in a JTM45 but have wired a few with shielded input wires and soldered the 68K resistors directly to the V1 tube socket but have since stopped this because it was really not doing anything for the amp.

[Steve_Hogan]

hi, I'm not an expert so I don't quite know what the correct answer is, but my understanding is that for the bias circuit you just basically need a voltage and there isn't much current draw (so seemingly no need for stringent ripple reduction as in lots of filtering, more stages, regulation and whatever). And the push-pull configuration maybe helps further by hum cancelling(?). As far as indiv. bias controls I think that is a very good idea in general (and if using old stock tubes especially nowdays with those only getting scarcer as time goes by).

Agreed on all points. It is true that the push-pull configuration will tend to cancel hum as long as the polarity of the ripple is the same at both grids. I have seen some dual bias proposals that use completely independent bias circuits but feed from opposite polarity windings. These won't cancel. If you use half wave, you must use only one side of the HT winding. If you do full wave, you get to use both sides and the ripple will cancel. Regulation would be a really bad idea IMHO due to reliability issues. The small 10uF caps cannot get much larger or they will take too long to charge up on turn on. The last thing you want to do is have a $400 matched pair of NOS GEC KT66 tubes begin to conduct on startup before the bias has stabilized.

In my spice simulation, the pp ripple on the bias voltage with the original half-wave bias circuit is 15.7mV (-33.86 dBu)
The pp ripple on the bias voltage with a full-wave rectifier (with revised circuit values as described above) is: 2.6mV (-49.48 dBu). That's a 15.62 dB reduction in hum simply by adding a second rectifier. That reduction is before whatever cancellation happens in the push-pull output stage. The fact that the bias reaches its ultimate value 2x faster with the full-wave than with the half wave gives added protection to the output tubes during warm-up. By the way the second diode can be implemented in a single bias as well.

[Steve_Hogan]

I was always under the impression that the bias circuit was such a low current application that it didn't require anything more than a single diode. Basically, it is not suitable to noise due to it's low current draw. Of course if one of the bias supply filter caps goes bad you will hear the hum come through for sure. If anything would cause some noise in the power section I would suspect of the tube rectifier 5V tap as this drops 5VAC directly onto the power rail that feeds the entire amp. With all this said, I haven't had any real issue with hum in a JTM45 but have wired a few with shielded input wires and soldered the 68K resistors directly to the V1 tube socket but have since stopped this because it was really not doing anything for the amp.

Because the 5V Rectifier Heater winding is floating, it doesn't add its own 5V to the Rectified DC. The shielded input wire would, I suspect, add additional HF rolloff due to increased capacitance to ground, and would probably make the amp sound a bit more dull.

[Steve_Hogan]

dai is too modest - he knows more than he is letting on!

You have to be careful when modding these amps. Remember it is far from hi-fi and the reason that they sound so good and give the tones that they do is that they are made of proprietary equipment available in the early 60's at a god price. Design having been borrowed from earlier Fender (and other designs).

The valve rectifier when working in conjunction with the slightly underspecced PT gives a significant sag in the power supply, so do not go for a bigger better modern PT and extra rectification - because more is better - it will not! same goes for the rest of the amp, fill it full of high spec mfs and polystyrene caps and you will have nothing like the organic BB tone you seek.

But yes it will cope with NOS GECs in that range!

-42.52V to -58.23V is the range of the original 1/2 wave circuit. I totally agree that "hi-fi" isn't what we are after, so the points you mention are already being addressed (hence the NOS mustard caps for example). The choice of resistor type is a careful balancing act between excess noise and getting the non-linear voltage dependent non-linearity that helps to make the sound "right". I believe, however, that all things being equal, lowering the hum and buzz of the amp cannot make it sound worse, so I am paying very careful attention to the grounding scheme and other easy improvements like the full-wave bias rectifiers.

[neikeel]

Not sure full wave rectification is needed but would be fascinated to hear a before and after.

Best resistors for balance/tone/ low noise in these are Piher 1/2w carbon films, IMO of course.

I think carbon comps add very little apart from noise, as you know they can soften things a little and can add a bit of vintage cracle and pop

[Roe]

didn't the major come with a full-wave bias circuit?

[Brian Wallace]

Yes it did. Good eye.