Bias worry

[bigwill]

I have made some new PCBs for someone's Radford MA25 amplifiers. The boards work, but unfortunately due to assumptions and miscalculations, I made the boards such that they bias the output valves at 60mA each (to within 0.1mA using an LM317 based CCS). Recently having the opportunity to measure his amplifier I discovered that his amp has some unusual component values and voltages - the cathode bias resistors are 680 ohms which drop about 35V, working out to be a current of almost dead on 50mA.

His pair are unusual, they have a GZ34 for rectification and are cathode biased, yet they are MA25s. Due to the GZ34 I wasn't expecting the HT voltage to be so high - it reaches about 420V, hence the need for high value cathode resistors (they are usually 390 or 470 ohms in the Radford series)

The 10mA extra per valve ends up drawing 10W more through the power transformer than normal, and although he doesn't have a preamp that these monoblocks can also power (and must suck quite a few watts), I'm quite concerned about this.

The reason I'm hesitant to just change the boards is due to the way I've implemented the bias system, I use 4 LM317s in parallel to bias one valve, so this means I'll have to change 16 resistors, potentially ruining the PCBs I've worked so hard on!

Will pulling an extra 10W through the power transformers be a huge problem?

Attached is a pic of the PCBs I've made for him. They implement the original radford circuit, plus a few extras consisting of a protection circuit and cathode fuses.

I don't have a wealth of experience with tube stuff being only 23 but I'm trying to do the best job I can since I'm being paid for these things, I really don't want to overheat his rare power transformer!

[analog_sa]

Zero experience with Radfords but i still wouldn't worry too much about the transformer. Can you estimate or measure the operating temperature?

Very curious why you needed so many 317s.

[SpreadSpectrum]

Nice board!

I know that the idea is intimidating, but you may want to consider incorporating some SMT resistors into the design next time. Then if you have to make a change it is a simple matter of finding some hot tweezers and a little solder wick with no real risk to the board.

If you decide to swap those resistors, risk of destroying the board is minimal if you don't keep heat on things too long and don't force anything.

Old farts don't like SMT because they are shaky and can't see, but us young guys can handle it. It really frees up a lot of space, too. I had an extra square inch on a board I recently did and added a USB DAC to get my money's worth.

[bigwill]

Originally Posted by analog_sa

Very curious why you needed so many 317s.

It's silly but I wanted to avoid messing with heatsinks as there's minimal height in the amp chassis to begin with, although of course if I were starting from scratch I would use one with a heatsink, which I have no problems agreeing with being better! Luckily with 4 in parallel they only run luke warm to the touch after running for hours, so they should survive. They have to shrug off about 0.25W, although this is the limit of what I'd subject a to220 package to without a heatsink!

Originally Posted by SpreadSpectrum

Nice board!

I know that the idea is intimidating, but you may want to consider incorporating some SMT resistors into the design next time. Then if you have to make a change it is a simple matter of finding some hot tweezers and a little solder wick with no real risk to the board.

If you decide to swap those resistors, risk of destroying the board is minimal if you don't keep heat on things too long and don't force anything.

Old farts don't like SMT because they are shaky and can't see, but us young guys can handle it. It really frees up a lot of space, too. I had an extra square inch on a board I recently did and added a USB DAC to get my money's worth.

SMD is tempting but I quite like the aesthetic of through hole components although they take up a lot of space and are a pain to remove!

If I were to do a DAC I would definitely consider SMD parts (Probably would be forced to in some cases) although I've never worked with them before apart from the time when I had fun pushing around little resistors (with a hot iron) on the back of a broken slot based Pentium 2 board

[ray_moth]

Old farts don't like SMT because they are shaky and can't see, but us young guys can handle it.

Being only 65 I can see the funny side of this - er - actually, I can't see. Sorry for this shaky response

[kavermei]

SMDs are surprisingly easy to work with once you try it out (if you have a steady hand, I suppose). Just stick to the "larger" form factors, nowadays there are Rs and Cs with sub-millimeter dimensions. Even IC's are easy, there are some tricks to do it right floating around on the web.