I'm just starting a build of Pete Millett's Mighty Midget compactron 6T10 amplifier and note that the Bill of Materials has a mix of metal film, carbon film and carbon comp resistors. I imagine Pete (being the wizard he is) has a good rationale for selecting the various types of resistors for this circuit.
Which resistors are critical for the signal path of the amplifier? There is a good sale on at Parts Connexion at the moment - some of the boutique resistors are over the top in terms or pricing but if anyone can recommed a reliable brand for these types it would be appreciated!
Which resistors are critical for the signal path of the amplifier? There is a good sale on at Parts Connexion at the moment - some of the boutique resistors are over the top in terms or pricing but if anyone can recommed a reliable brand for these types it would be appreciated!
Mainly keep the grid resistors as carbon composition types, in the same wattage as spec.
The others can be anything but carbon comp. Standard metal film or carbon film is fine.
Expensive resistors are not needed. Keep the same wattage as spec to keep the price and size down.
Higher power resistors may have to be metal oxide types depending on stocking.
Don't change the resistor wattage except higher. The R3 should be wirewound because of surge current.
The others can be anything but carbon comp. Standard metal film or carbon film is fine.
Expensive resistors are not needed. Keep the same wattage as spec to keep the price and size down.
Higher power resistors may have to be metal oxide types depending on stocking.
Don't change the resistor wattage except higher. The R3 should be wirewound because of surge current.
Last edited:
Okay great, appreciated - which particular resistors in the circuit would be important in terms of reducing the noise floor (if any) of the amplifier?
A very interesting set of questions, and discussion.
Generalizations:
Most quality (and intact) resistors have noise power according to the rule KTB.
T is the temperature. Higher temperature = higher noise power
Higher resistance means higher voltage, but lower current.
Lower resistance means lower voltage, but higher current.
What happens when a resistor has Either DC or Large Signals going through it?
It causes a rise in temperature, T; which sounds like . . . more noise.
Other problems that some resistors have, is that they are not properly constructed.
If the end caps do not make low resistance and consistent contact.
If a batch of carbon is loose or not evenly packed.
Oxidization of the metal contacts.
Other mechanical and chemical problems Etc.
DC current potentially causes noise to be generated by bad end caps, etc.
What resistors have DC passing through them, but do Not have any capacitor shunting their noise to ground?
R16 and R17
R11 and R10
R18 and R20
Just saying.
Generalizations:
Most quality (and intact) resistors have noise power according to the rule KTB.
T is the temperature. Higher temperature = higher noise power
Higher resistance means higher voltage, but lower current.
Lower resistance means lower voltage, but higher current.
What happens when a resistor has Either DC or Large Signals going through it?
It causes a rise in temperature, T; which sounds like . . . more noise.
Other problems that some resistors have, is that they are not properly constructed.
If the end caps do not make low resistance and consistent contact.
If a batch of carbon is loose or not evenly packed.
Oxidization of the metal contacts.
Other mechanical and chemical problems Etc.
DC current potentially causes noise to be generated by bad end caps, etc.
What resistors have DC passing through them, but do Not have any capacitor shunting their noise to ground?
R16 and R17
R11 and R10
R18 and R20
Just saying.
Last edited:
Note that chassis grounding is not done correctly in this drawing. PE (the safety Earth connection to wall outlet) must be connected directly to chassis, with a dedicated bolt. Signal ground can (well, kinda/sorta/maybe can) be lifted from chassis with a 35A diode pair or bridge, which should have a small (10R is fine) resistor bridging it to swamp residual leakages and therefore keeping normal operating voltage between signal ground and chassis near zero. Some folk also parallel a 0u1F capacitor, but that's application specific, a judgement call, and not safety related.
All good fortune,
Chris
All good fortune,
Chris
Last edited:
R10, R11, R12 and R13 are the grid resistors.
I can't locate any carbon comp resistors in the range of 500R to 2KR rated for 1/2W.
What is the next best alternative for these - metal film?
Rayma,
Agreed.
R10 1k, and R11 1k Ohm resistors are relatively low resistance.
The noise power of a lower resistance is larger current, and less voltage, versus noise power of a high resistance.
The output stage is either in Pentode, or in Beam Power mode. The output tubes are using Schade negative feedback.
But the signal current through them can be fairly large, when the signal pulls the plate voltage near to the cathode voltage.
That means the screen voltage is much higher than the plate voltage, and so they draw plenty of current at the peak of the signal.
Agreed.
R10 1k, and R11 1k Ohm resistors are relatively low resistance.
The noise power of a lower resistance is larger current, and less voltage, versus noise power of a high resistance.
The output stage is either in Pentode, or in Beam Power mode. The output tubes are using Schade negative feedback.
But the signal current through them can be fairly large, when the signal pulls the plate voltage near to the cathode voltage.
That means the screen voltage is much higher than the plate voltage, and so they draw plenty of current at the peak of the signal.
Noise: there are two (broad) categories of noise in resistors, thermal/Johnson/Nyquist noise which is completely predictable, depending only on the value of resistance (in Ohms) and temperature (in K), and every other noise (so it's called "excess noise"). Excess noises tend to be process sensitive, sometimes current or temperature sensitive, in bad cases voltage sensitive. Nothing you can do about thermal noise, but you can choose among resistor types for least excess noise, best long-term survivability with (especially voltage) overloads, and (for grid stoppers) minimal parasitic reactances.
Distortion: measurable in very bad cases, with great difficulty, but in general, resistors with low excess, especially 1/f, noise tend to also be low distortion. These may be related, but probably nobody understands the mechanism. It's actually cutting edge science, who knew?
Critical?: For noise everything scales in importance with its relation to signal level - the smaller the signal at that point, the more important the noise contribution, but generally the least exciting of excess noise mechanisms. For distortion everything scales in proportion to expected distortion at that point. Perhaps strangely, it's possible for the first stage grid leak resistor to contribute more distortion to the total than any other resistor (but really only in a truly pathological case).
All good fortune,
Chris
1. If you have too much noise coming out of your amplifier, look at these 4 things first:
Noisy Tube
Bad resistor
Amplifier topology; too much gain; etc.
Oscillations; bursting, interference from wiring, power supply hash, ground loops, etc.
If that does not fix it, you have some real serious troubleshooting ahead.
Claims about carbon comp resistors should be considered in the light of logic:
2. A small resistor (low wattage) that is constructed by wrapping the resistive element around ceramic, etc., and that has end caps, can be highly reactive at RF frequencies (resistive wrap and leads are inductive, end caps are capacitive to each other).
A larger resistor, a 2 Watt carbon comp for example, mainly only has inductance from the lead wires, the capacitance of the leads at the carbon element is minimal.
Hint: guess why the grid stopper resistor is properly connected directly (closely) to the tube socket tab. Reduction of inductance? Yes.
Next subject . . .
Your Mileage May Vary
Noisy Tube
Bad resistor
Amplifier topology; too much gain; etc.
Oscillations; bursting, interference from wiring, power supply hash, ground loops, etc.
If that does not fix it, you have some real serious troubleshooting ahead.
Claims about carbon comp resistors should be considered in the light of logic:
2. A small resistor (low wattage) that is constructed by wrapping the resistive element around ceramic, etc., and that has end caps, can be highly reactive at RF frequencies (resistive wrap and leads are inductive, end caps are capacitive to each other).
A larger resistor, a 2 Watt carbon comp for example, mainly only has inductance from the lead wires, the capacitance of the leads at the carbon element is minimal.
Hint: guess why the grid stopper resistor is properly connected directly (closely) to the tube socket tab. Reduction of inductance? Yes.
Next subject . . .
Your Mileage May Vary
Last edited:
Bulk material resistors like CC have the lowest inductance, same as a wire.
Must use CC types. The rating can be 1/4 W, there's no significant power dissipation.
Or can use 2k 1/2W also. Some use as high as 4.7k.
Ah finally some numbers. https://www.vishay.com/docs/31023/erl.pdf
10nH inductance for this series of metal film resistors.
10nH inductance for this series of metal film resistors.
The scarcity of carbon composition resistors is rather alarming!
I can source some 3.3k 1/2W 350V carbon comp types for the grid resistors. What does increasing the value of the grid resistors actually 'do' in the circuit (noting the designer has specified 500R to 2kR as the range for R10->R13)?
I can source some 3.3k 1/2W 350V carbon comp types for the grid resistors. What does increasing the value of the grid resistors actually 'do' in the circuit (noting the designer has specified 500R to 2kR as the range for R10->R13)?
Thanks @rayma that's excellent!
On another note, what is the modern thinking on volume pots? In the audio chain these days, there is often numerous other means of controlling volume (whether in the digital realm or analogue stage), so is having a passive ALPS style volume pot such as in this design entirely necessary or does it degrade the overall sound of the circuit? I guess the question is, would the circuit be "purer" if the pot were excluded? I'd be relying on other people's experience of using pots versus not using them when volume is typically controlled in earlier stages of the audio chain.
Another alternative might be a resistor step style attenuator but I don't think there are any pin-compatible ones around that would fit the PCB for this particular circuit?
On another note, what is the modern thinking on volume pots? In the audio chain these days, there is often numerous other means of controlling volume (whether in the digital realm or analogue stage), so is having a passive ALPS style volume pot such as in this design entirely necessary or does it degrade the overall sound of the circuit? I guess the question is, would the circuit be "purer" if the pot were excluded? I'd be relying on other people's experience of using pots versus not using them when volume is typically controlled in earlier stages of the audio chain.
Another alternative might be a resistor step style attenuator but I don't think there are any pin-compatible ones around that would fit the PCB for this particular circuit?