Wirewound is good for most audio amplifiers, but some circuits require low inductance emitter resistors for stability.
Circuits with local feedback loops closed around the output devices (CFP), with MOSFETs, or with very high unity-gain bandwidth products (>4Mhz) are likely to oscillate when wire-wound resistors are used.
Note that the oscillation may show only in certain circumstances, for example near clipping, so careful testing is recommended.
Circuits with local feedback loops closed around the output devices (CFP), with MOSFETs, or with very high unity-gain bandwidth products (>4Mhz) are likely to oscillate when wire-wound resistors are used.
Note that the oscillation may show only in certain circumstances, for example near clipping, so careful testing is recommended.
Eva said:
That sounds like a good idea, I'll try that. Also, what about a capacitor across it?
I have a question, being there's a lot of mixed views in here about the inductance of the wirewounds, and also you mentioned CFP.
My subwoofer amp I just built is CFP, but I'm using ceramic wirewounds because they are what I had, and are 0.15 ohm 10W which is perfect for a big amp with a 4 or 2 ohm load. Now while building this amp, I noticed I had stability issues when I didn't use resistors at all, but I changed that along with a few other things, and the amp seems really stable with the resistors. Haven't had any extra heat, or any biasing problems, and have run it to +/- 52V so for just testing.
Should I worry about having a CFP with ceramic wirewound output resistors, or is the low 0.15 ohm rating not going to make much inductance to worry about? Also, for the 10 outputs, they each have a compact, low inductance, wirewound 0.13 ohm 3W resistor, that's the size of a 1W film resistor.
Attachments
AndrewT said:
The dual VBE is because I am monitoring 2 sinks, one for NPNs and one for PNPs. The bias doesn't change much once set, and is rather stable, hot or cold. I tested the amp at
350mA quiescent with +/- 52V and the sinks get hot, but the bias never runs away, or overcompensates. So far so good. I'll run it probably at around 100mA or less when finished.Also, the extra 1 ohm resistor and 10nf cap in my Zobel is to bypass any other RF or HF signals that the other zobel may have missed. I want to make sure this thing never oscillates.
I have an inductance meater that reads well below 1uH. I tried it out on a range of resistors to satify my curiosity. Here is some of what I found.
- Wirewound < .47-ohms : Less (usually far less) than 1uH. Some so low they couldn't be measured especially .22-ohms or less.
- Wirewound > .47-ohms to 10-ohms: Between 10uH and 100uH. Definately inductive.
Obviously for wirewounds, linductance varies with resistance. Number of wraps? No big surprise. Typical Re values (<.47 ohms) can probably be considered "virtually non-ductive". Nonetheless, I would urge you to measure them case by case if you have the means. Some manufactures datasheets (*if* you can find them!) mention inductance but usually not or least not in much detail.
Metal film power resistors:
Be careful about your assumptions!!
Below a particular value, typically 200--300 ohms many are inductive. This is sometimes specificly mentioned in the data sheets. The reson is simple-- for a given series, below a particular value the manufacture cuts a spiral into the deposited metal film in order to attain the desired resistance value. This makes the metal film resistor functonally equivalent to a wirewound resistor. DO NOT assume a metal film resistor is non-inductive in all cases. Some are significantly inductive. Measure for yourself if possible. If not consult a datasheet.
The range with the greatest ambiguity is >.33 --< 400. Lower, the inductance is normally barely measurable for any resistor, and above this range metal film resistors are normally non-inductive. In between requires more caution.
Don't forget zobel resistors or the damping resistor parralel to an output inductor. These are typically 5.6 - 10 ohms and may be critical for MOSFET output devices.
The only convenient , reliable (meaning I know unambiguously what I'm buying) source I've found when I need a low value non-inductive power resistor is Mills non-inductive resistors that I order from Michael Percy.
Hi EWorkshop,
nice idea a super vhf Thiel network. To make it effective the routing on the PCB MUST be low inductance (i.e. very short track lengths) and following those other replies the resistor must also be non inductive. A secondary function and some would say the most important function of the Thiel network is to attenuate RF signal coming back in through the output leads and entering the NFB loop striaght into the LTP amp. A second Zobel after the output inductor would assist this.
nice idea a super vhf Thiel network. To make it effective the routing on the PCB MUST be low inductance (i.e. very short track lengths) and following those other replies the resistor must also be non inductive. A secondary function and some would say the most important function of the Thiel network is to attenuate RF signal coming back in through the output leads and entering the NFB loop striaght into the LTP amp. A second Zobel after the output inductor would assist this.
Mr Evil said:
These are good for the output Zoebel, as the Zoebel is intended to shunt RF. Low inductance there is good, and their noise is not an issue. I wouldn't use them for output emitters because of their really poor temperature stability. I've used them to wind the output inductor around them, too, because they are nice and cylindrical, and don't have any magnetic material to be inside the inductor.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.