Thanks for the suggestion Nigel. I think this is going to be a frustrating one to track down.
I took the lid off tonight and, following your suggestion, found that sometimes the unit is incredibly sensitive to the position of the secondary outputs from the transformer (I had to weave the wires between the reservoir caps to make the hum go away), yet other times the positioning of the wires makes no difference and no hum is heard regardless of position. Powering it off and then back on again can sometimes force a change in this behaviour, but earlier I tapped the front of the unit out of frustration and the hum gradually faded out over a few seconds, maybe suggesting a mechanical issue or bad component.
When it is misbehaving the hum is audible on both channels even with no source connected. The channels are each being fed from their own secondaries wound on the same transformer.
Last edited:
Interestingly, the quoted Beta values are towards opposite ends of the C suffix range of 420 to 800.
While the differences in the Betas are 3-4 percent, looking at Alpha (Ic/Ie) = Beta/(Beta+1) shows a closer relationship between these internal currents.
For a Beta of 437 and half the Nait 3 LTP current of 0.5 mA 437/438ths of 0.5 mA is 0.4988mA. - the partnering TR with a Beta of 455 giving 455/456ths of 0.5 mA or 0.4989 mA.
In the other channel the equivalent collector currents are 0.49927 mA and 0.49929 mA.
You should be able to emulate this level of matching with your selections.
Another way of matching is of base currents.
Dry joint, bad joint or bad component. Could be a bad capacitor. If you had a scope you could check various test points.
Is the hum 60 or 120Hz? (50 or 100Hz for other parts of the globe).
EDIT: Scrap that. I see it is both channels. I had same problem before on one of my amps, but it was because my signal GND was not connected to my main audio GND. I doubt you would make this mistake.
Last edited:
Does this affect your opinion as to whether a pair of matched BC550C would be a suitable replacement for the MPSA06?
It is an odd one, not sure why both channels are affected.
I have got a DSO, I'll do some further digging when I get chance. I did a quick check on the speaker outputs last night and it is 50Hz.
As always it's frustrating due to it being intermittent!
1) Check all 0V connections for continuity and low impedance
2) If you are using phono cables on the NAP input, twist them together to minimise the loop area between the L & R
3) Have a sniff around with some freezer spray
4) Short both inputs and see if you still get the hum
Probably best not to use half decent speakers when investigating.
2) If you are using phono cables on the NAP input, twist them together to minimise the loop area between the L & R
3) Have a sniff around with some freezer spray
4) Short both inputs and see if you still get the hum
Probably best not to use half decent speakers when investigating.
I'm starting to run out of T3.15A fuses...
I'll build something similar to your lamp-in-series solution, Nigel, in the short term (I've been using a series bulb for testing purposes anyway), but I've just ordered one of the following that I'm expecting will be a permanent solution that will also protect against the occasion when the mains is restored following a power failure:
220V 1000W Transformer Delay Power Soft Start Protection Board for Amplifier AMP | eBay
Chris
Sometimes it's outside the house. Often mechanical noise is heard at the same time. Measure between Neutral and Earth on AC range and see if it relates. Could be 0.2V or if at our lab 6V! We had a clean supply for special work.
The soft start ciruciit looks good value. I suspect it is a little bit naughty as I see no mains transformer. It could be a heatless balast. If so the capacitors should be class X2 or better ( even then I doubt it is allowed as it is playing tricks to get it to work ). These look like standard high voltage caps with no X2 classification. I could be very wrong so don't take my word for it.
Last edited:
This sounds like a LOOP AREA problem.
Any wire pair that flows and returns current from/to a Source must use low loop area to minimise interference when the current changes.
If you don't use twisted triplets for your transformer to rectifier and from rectifier to smoothing capacitors and from smoothing capacitors to amp PCB, then you are increasing the interference inside your amplifier.
This requires a 4 secondary transformer.
If you have a dual secondary transformer and simply double up the bridge rectifiers to 4 instead of the normal 2, this does not isolate the channels. You will be creating extra LOOPS
Last edited:
The main secondaries are both 28-0-28, there are two of these independently wound on the same core.
There's a third, which is 26-0-26 for the pre-amp PSU section.
This isn't a problem all of the time, but can be made to appear and then vanish by switching the mains off and on. It's very odd. The hum disappears immediately when disconnecting the mains power.
Last edited:
That requires YOU to take safety precautions to prevent Mains getting to other components/fingers inside the enclosure.I'm starting to run out of T3.15A fuses...
I'll build something similar to your lamp-in-series solution, Nigel, in the short term (I've been using a series bulb for testing purposes anyway), but I've just ordered one of the following that I'm expecting will be a permanent solution that will also protect against the occasion when the mains is restored following a power failure:
220V 1000W Transformer Delay Power Soft Start Protection Board for Amplifier AMP | eBay
Chris
It is very similar to the soft start offered by the Forum and I have criticised it for breaking our Forum rule on no direct mains connections.
The hum on my system seems to come from the outside next to my a firm has a compressor and each time it starts I have what is better described as a train going from left to right not un like Doppler shift. Note I am a brit living in Brazil the local supply is 220v 3ph and 120 single phase and no ground interesting to say the least. And the compressor is across two phases and like all big single phase motors is cap start. One last point is it been picked up in the speaker cables mine did pick up some. Measured the output with speakers connected 0.018 mv and with dummy load connected 0.07mv this was tested with a new Fluke 177 fresh from calibration
The last confirms it is a Loop problem.
Mains current interference is getting into an amplifier circuit.
OK thanks, I'll focus efforts on resolving this from that perspective then.
I metioned the transformer rotation as an Exposure amp foxed myself and it's PHD in electrical engineering friend ( His BSc was in electronics, his amplifier he couldn't fix). I went through everything and then found I could null it completely if the transformer was vertical. This was out of the question so the least worse position chosen. There were two options. Statistical lowest and ears lowest. The ears lowest was a nearly single higher value lower frequency scope readout. With the spare windings it might just be more fussy.
What fooled us is Exposure had set the position with a two bolt fixing. This was obviously right when prototype, but was not checked in production. New bolt holes made to refix. As luck would have it just possible to get a result.
Exposure had a twin regulator design ( 4 in total ). This had two goes at curing the problem. This to my mind is why Exposure never beat Naim in the market place.
The fix was well worth doing. The sound was deeper and had more clarity. What is silly is I took two weeks to find it. I just trusted what they had done.
Whilst it is unlikelysomething so simple it is worth checking. You don't need a scope. Colin Wonfor pointed this out to me years ago. If fitting an MC stage to his amp he suggested nulling the transformer. Thanks Colin.
What fooled us is Exposure had set the position with a two bolt fixing. This was obviously right when prototype, but was not checked in production. New bolt holes made to refix. As luck would have it just possible to get a result.
Exposure had a twin regulator design ( 4 in total ). This had two goes at curing the problem. This to my mind is why Exposure never beat Naim in the market place.
The fix was well worth doing. The sound was deeper and had more clarity. What is silly is I took two weeks to find it. I just trusted what they had done.
Whilst it is unlikelysomething so simple it is worth checking. You don't need a scope. Colin Wonfor pointed this out to me years ago. If fitting an MC stage to his amp he suggested nulling the transformer. Thanks Colin.
Why? The Forum think it's OK to break their own rule !
I have asked both privately and on the Forum for a clarification and/or re-wording to show what we are allowed to discuss. They have never come back to me with any new safety rules/suggestions.
The rule is completely ignored in all the Threads discussing mains input SMPS.
Can you imagine an electric toaster being allowed were it not for the past.
I wonder if we really should be allowed to wire on a new plug? We are, it's a very reasonable question whether we should?
Out side of the UK it might intrigue people to know our grandmothers very much could and can do this. Not only that, but do it correctly. There stil is on UK plugs a guide as to how to do it. It involves making live shorter as first step to safety. The screws on the connections are flat to avoid damaging the wire. Some imported pugs got that wrong. The UK plug is a very good device.
Brass was to be forbidden as it contains lead typically. It could not meet lead free regulations. It took some careful arguement to get the powers to realise the lead was safely locked up when brass. If they had persisted life as we knew it would have been impossible. The water supply also uses it.
Lead free is not true for aircraft as it is considered a safety risk to use it. It is also certain more lead exists on church roofs than in waste sites due to PCB's. Acid rain does cause genuine problems with lead roofs.
I wonder if we really should be allowed to wire on a new plug? We are, it's a very reasonable question whether we should?
Out side of the UK it might intrigue people to know our grandmothers very much could and can do this. Not only that, but do it correctly. There stil is on UK plugs a guide as to how to do it. It involves making live shorter as first step to safety. The screws on the connections are flat to avoid damaging the wire. Some imported pugs got that wrong. The UK plug is a very good device.
Brass was to be forbidden as it contains lead typically. It could not meet lead free regulations. It took some careful arguement to get the powers to realise the lead was safely locked up when brass. If they had persisted life as we knew it would have been impossible. The water supply also uses it.
Lead free is not true for aircraft as it is considered a safety risk to use it. It is also certain more lead exists on church roofs than in waste sites due to PCB's. Acid rain does cause genuine problems with lead roofs.
I had this kind of intermittent line noise in previous circuits, mainly high gain preamp. It was always ground related.
Turning the transformer for minimum noise is also sometime necessary. YBA was doing it on their amp assembly line. I had to do it some time in tube preamp or other high impedance circuit. It may worth a try.
Are you using a Naim preamp with it? Naim has a peculiar gnd in their system where all the gnd are brought back to the preamp I think. It was explained somewhere else in this thread.
Turning the transformer for minimum noise is also sometime necessary. YBA was doing it on their amp assembly line. I had to do it some time in tube preamp or other high impedance circuit. It may worth a try.
Are you using a Naim preamp with it? Naim has a peculiar gnd in their system where all the gnd are brought back to the preamp I think. It was explained somewhere else in this thread.
Right, problem solved.
The cause? A faulty bias trimmer in one of the channels!
I identified there was a fault by finding that the voltage across the 2 x 0R22 resistors was ~2.5mV. Knowing that the voltage here was set to ~13mV during initial setup I probed further.
When in-circuit the resistance across the 1K resistor connected across the trimmer was always 1K regardless of the trimmer setting. However when I pressed a little harder whilst adjusting the trimmer the resistance changed erratically. Could have been a poor solder joint so I removed it to test it.
Out-of circuit the trimmer tested fine. Put it back in, same erratic result.
Replaced it (and the one in the other channel just in case) and now all is well.
Thanks for all the suggestions, it's been a learning experience.