Constant Current Diode?

[thermionic]

Hi,

A power amp I have uses a 1.5mA / 100v 'constant current diode' to regulate current to the collectors of the LTP input transistors. I appreciate that constant current diodes can be referred to by several different names, but I can't seem to find an equivalent part at any of my suppliers... The diodes themselves look like zeners; a red body with black cathode marking.

Can anyone give me some tips as to what I should be searching for? I've checked this link out and searched for all the terms @ Farnell / RS etc: http://en.wikipedia.org/wiki/Constant_current_diode They only seem to carry a handful of TO-92 diodes with different current ratings and fearsome price tags. I guess it won't hurt if I fit a diode with a larger rating, say 2.2mA. The idea of the diode is to regulate current to the LTP without any loading from the OPS affecting them. I have a separate CCS for the current mirror. I'd love to post the schematic, but it's copyrighted and I had a bit of a falling out with the designer, so I'd prefer not to give him opportunity to give me grief. It's a fairly basic pwr amp, and I doubt any of you will lose sleep not seeing it.

Thanks in advance.

Justin

 

[Svart]

1N5301
1.53mA
100v

Constant current diodes are actually a JFET in a diode package..

 

[thermionic]

Thanks, Svart! I just found this: http://www.centralsemi.com/PDFs/products/1N5283-5314.pdf

Justin

 

[Samuel Groner]

I guess it won't hurt if I fit a diode with a larger rating, say 2.2 mA.

Increasing the bias current of the input pair may seriously mess with the designers intention--although a step from 1.5 mA to 2.2 mA should probably be handled reasonably well.

I doubt you really need a 100 V rating unless this is a really high power amp. Linear Systems carries parts with 50 V IIRC.

If you don't find a suitable part you can use a JFET or depletion mode MOSFET (e.g. DN2540) and select a source resistor to give 1.5 mA.

Samuel

 

[Svart]

and I forgot to mention that Mouser carries them.

 

[thermionic]

The amp has 75v rails, and the diode goes straight to +75v, above the LTP's collectors.

I've found a supplier in the UK for the Mouser-listed part, but I've read that a jfet + R current source works better... What advantage would this give me? Is there any point, aside from learning how to make a current source, in my app?

Thanks to both of you.

Justin

 

[JohnRoberts]

Well if you add stray capacitance there, you could learn more about replacing parts if the amp starts oscillating.

If you roll your own current source maybe add a series resistance to decouple out any new extra capacitance.

Lower noise there couldn't hurt, while it may not be the limiting factor. Adding noise there could be audible.

Does sound like a learning experience... When repairing power amps, we used load boxes that current limited the mains power while using a variac to bring up the mains voltage.

Oops with power amps can release even more of the magic smoke...

JR

 

[thermionic]

Well if you add stray capacitance there, you could learn more about replacing parts if the amp starts oscillating.

If you roll your own current source maybe add a series resistance to decouple out any new extra capacitance.

Lower noise there couldn't hurt, while it may not be the limiting factor. Adding noise there could be audible.

Does sound like a learning experience... When repairing power amps, we used load boxes that current limited the mains power while using a variac to bring up the mains voltage.

Oops with power amps can release even more of the magic smoke...

JR

This design of amp has oscillated in the past...(going back years). I can't remember exactly how I cured it, but I seem to remember replacing the Zobel parts (Zobel goes to GND, just prior to the inductor / parallel damping resistor @ output)

What's more noisy: a CCD or discrete jfet / R CS?  

If the amp's relatively stable with a device like that which Svart recommended, wouldn't I be best to let sleeping dogs lie?

I know all about smoking amps...

Thanks

Justin

btw John - maybe you need a replacement smoke kit

 

[Samuel Groner]

The diode goes straight to +75 V.

You could put a zener in series with it--will also lower the dissipation.

I've read that a JFET + R current source works better... What advantage would this give me?

Typically the output resistance is slightly higher with added source resistor. Also the temperature coefficient might be lower, although I'd need to look at this again to be sure. In any case rather small effects for your application.

Well if you add stray capacitance there, you could learn more about replacing parts if the amp starts oscillating.

This is not usually a sensitive node regarding stability, at least as long as strays are referred to ground. Added stray capacitance will surely worsen HF CMRR and perhaps PSRR though, but typical amplifier designs tend to have other serious restrictions with this respect so the effect might not even be noticeable in the overal picture.

Adding noise there could be audible.

You need a hell lot of current noise till there is any effect within the audio frequency range because any noise appears as common-mode signal which is rejected by the amplifier CMRR. JFET CCSs tend to be very low noise anyway.

Samuel

 

[JohnRoberts]

Yes in an ideal world it's all common mode,, 

I have found in power amps because of the high current flowing in output stages power supply rejection issues "can" be important.

Assuming it is a good stable design, there will probably not be an issue.

Go for it..  if it oscillates call Sam..  ;D

JR