AMC 3030 annoying ripple noise

[SimpleMind]

Dear All

I recently got an 1982 AMC 3030 hybrid amp. The advertised condition was: everything running and fine. The condition in real world terms was a bit different: severe hum (50 Hz cycling of our European power lines) and one channel went dead after a couple of hourse listening.

As I loved, what I heared from this amp in the mean time (through my Technics SB-660), I decided to try a repair and at the same time to add some modifications, which I wanted to have anyway.

Schematics can be found on the official web page: AMC CVT3030MKII - CCVT Integrated Amplifier, Class A, 30W x 2, with Phono Input Stage | AMC (scroll down, the service manual is near the bottom of the page).

When I opened the amp, I couldn't trust my eyes: repair traces everywhere. One tube was replaced, wires added, where the pcb tracks were broken etc.

I did the following things:
-> desoldered the old tubes and soldered sockets into place, to facilitate tube swapping
-> discovered a rectifier diode, which made no contact to the pcb
-> replaced all rectifier diodes with higher voltage, UF types
-> replaced the higher wattage resistors with new ones, partially with even higher wattage
-> replaced the bridge recitifiers and the voltage regulators (7915 + 7815)
-> replaced most electrolytics, incl. the main filtering cap

-> added RCA sockets for pre-amp out and power-amp in and used highly shielded cables for wiring these connections (still can use the internal pre-amp section by using a bridge cable)
-> replaced Z-diodes on the power amp modules
-> some resoldering and repairing of defective copper tracks on the pcb
-> adjusted BIAS and balance
-> replaced all the trim pots on the tube modules, because BIAS was not only erratic, but oscillated heavily and was very hard to adjust, as the tiniest movement on the pots would send the Bias from -6V to perhaps +120V

Interestingly enough, the adjustments with the new, expensive Cermet pots is in no way better, than with the old ones...

Result: The amp plays beautifully with new EL34 tubes (Svetlana S) and 50 Hz hum was gone. Instead now I have a ripple noise, something like 100 Hz from some rectifier? The noise is constant and independent of the volume. It stay the same, even when any pre-amp (including the AMC's own one, is turned down completely.

I had a look with the oscilloscope and found those typical (well, typical compared with examples I found on the web) spikes, produced by the rectifier diodes. May be that is the reason?

So I read up a lot about filtering out these spikes by adding snubbers (R 100 Ohms-C 68mF) across each diode. The noise got a bit quieter, but far from being gone.

Then I made a stupid mistake. While measuring, I disconnected the tube section and without thinking I reconnected the plug between the power supply board and the pre-amp board, while the amp was still switched on. The surge made nice sparks... No need to say, the fuses where blown and blew again, when I switched on the amp again... Lots of searching led to the following:
-> rectifier diode on the pre-amp board replaced, as one was blown
-> replaced Z-diodes on the pre-amp board
-> searched long to get the JFets J105, to replace the dead ones on the pre-amp boards

There is still a blown part in the headphone amp, but I simply disconnected it and now the amp runs nicely again. But this did not help to reduce the very prominent constant ripple noise. I would even think, it got worse, but that may have to do with my diminishing patience...

As it is now, I am lacking ideas and experience to trace the problem. What I see on the oscilloscope is only a very low amount of noise (perhaps like 10-20 mV) - but I can hear it very loud from the speakers.

What would be the cause:
- power supply? What else can I do. Perhaps the snubbers are completely wrongly dimensioned and are part of the problem?
- or adding snubbers across the bridge rectifiers, too?

- or is a problem with the tube modules? Bias wrong?

- As I can hardly see anything meaningful on the oscilloscope it would be hard to trace the noise across the boards to the cause of it all.

I would love to finish my work on this amp, as it already cost me a couple of days and quite some money for the exotic transistors and other spare parts. And also, I love the sound, the amp produces, just not the noise...

Thanks for reading and thanks for any hint into the right direction!

Regards
Ben

 

[AllenB]

It's interesting you said you found a rectifier that was not soldered and when you repaired this the hum went from 50Hz to 100Hz. Can you measure ripple on the supply rails? Are the supply filters original?

 

[SimpleMind]

It's interesting you said you found a rectifier that was not soldered and when you repaired this the hum went from 50Hz to 100Hz. Can you measure ripple on the supply rails? Are the supply filters original?

Hi Allen
Thank you for your reply. Well, yes, good you mention the not-really-connected rectifier diode. This might be the reason, that the hum changed its base frequency. In terms of supply filter, there is not much.. a couple of smaller (10µF to 47µF) caps and one bigger 390µF/450V cap. I replaced all of them with higher temperature and voltage types, as far as possible. The main cap was replaced with a nice Epcos 470µF/450V cap.

There is one thing, which I forgot to mention: the AMC has got two cooling fans. And one of the first thoughts I had, was, that they could be the reason for the noise. But I tried the amp with the fans unplugged, which didn't have any positive effect.

I will try to trace the ripple, but have a question: How can I visualize the ripple on the ground rails? I am not really oscilloscope-savvy but will happily learn. Where would I connect the ground of the osci's probe, when I try to measure the ground rail with its tip?

Another thing, I wonder about, is whether all the ripple and noise could come not from the power supply, but from the tubes directly. AMC gives a BIAS of 6V for the EL34. When I replaced the tubes and the BIAS voltage was way too high initially, they produced lots of noise, which went massively down, when I also brought down the voltage to the given value. But may be, there is still some lee way for improvement?

Ben

 

[jerluwoo]

The change in noise frequency means you have went from half-wave rectification to full-wave as it should be. There is room for improvement in the B+ and bias supplies, based on what the schematic shows there very little actual filtering going on. Starting with the bias, you can insert a 47k resistor between the pots and ground to reduce the adjustment range so it won't be so touchy. If you can't get the bias up to +6 volts then this value can be reduced. In the bias supply itself, replace C402/403/404 all with 100uF 63V,remember to observe polarity, insert 100ohm resistors between C402-403 and between C403-404. This should help clean up the bias supply a fair amount. The B+ you can add a 10 watt 10-22ohm resistor after C401 then another 330-470uF cap after that. Larger values of resistor will give more ripple reduction but trades more B+ voltage reducing power output. This may be fine though since the difference in volume between 25 and 30 watts is very little. Also not thrilled about the way the fans are powered, using the negative regulator for the preamp section, a total bean counter move. Perhaps adding a provision to power them from an external 12V wall wart would be better.

 

[AllenB]

I will try to trace the ripple, but have a question: How can I visualize the ripple on the ground rails? I am not really oscilloscope-savvy but will happily learn. Where would I connect the ground of the osci's probe, when I try to measure the ground rail with its tip?

Are you wanting to measure each end of the buss relative to each other or do you just want to know what is on ground?

Typically you assume that ground will be the quiet point and you don't measure it, but you measure everything from it. Everything is relative and even if ground were moving relative to some point in the universe, it's how the rest of the circuit moves relative to ground that is of interest.

When I replaced the tubes and the BIAS voltage was way too high initially, they produced lots of noise, which went massively down, when I also brought down the voltage to the given value

With more current flowing your power supply would have been under more stress. Not all supplies would give a noticeable result like this but it's possible.

 

[SimpleMind]

The change in noise frequency means you have went from half-wave rectification to full-wave as it should be. There is room for improvement in the B+ and bias supplies, based on what the schematic shows there very little actual filtering going on. Starting with the bias, you can insert a 47k resistor between the pots and ground to reduce the adjustment range so it won't be so touchy. If you can't get the bias up to +6 volts then this value can be reduced. In the bias supply itself, replace C402/403/404 all with 100uF 63V,remember to observe polarity, insert 100ohm resistors between C402-403 and between C403-404. This should help clean up the bias supply a fair amount. The B+ you can add a 10 watt 10-22ohm resistor after C401 then another 330-470uF cap after that. Larger values of resistor will give more ripple reduction but trades more B+ voltage reducing power output. This may be fine though since the difference in volume between 25 and 30 watts is very little. Also not thrilled about the way the fans are powered, using the negative regulator for the preamp section, a total bean counter move. Perhaps adding a provision to power them from an external 12V wall wart would be better.

Hi Jeremy

Thanks so much for your suggestions. I'll try your recommendations asap, though I need the high wattage resistors first... Yes, the fans make me feel nervous as well. But on the other hand, if I unplug them, the ripple noise won't change at all. Nevertheless powering them independently from the audio circuit is surely a good move.
A couple of watts more or less in power output is not relevant to me, as the amp is meant to power fairly efficient speakers anyway (currently the Technics SB-660 and later my homebuilt horn speakers).

Best regards
Ben

 

[SimpleMind]

Are you wanting to measure each end of the buss relative to each other or do you just want to know what is on ground?

Typically you assume that ground will be the quiet point and you don't measure it, but you measure everything from it. Everything is relative and even if ground were moving relative to some point in the universe, it's how the rest of the circuit moves relative to ground that is of interest.

With more current flowing your power supply would have been under more stress. Not all supplies would give a noticeable result like this but it's possible.

Hi Allen

Thanks a lot for the follow-up and the vivid clarification. I'll do a trace and try to find out, at which position on the different PCBs the noise really origins from - if there is single source. And I'll also try Jeremy's modifications. I'll try to capture the osci's screens, perhaps somebody will profit in future, if I document the work on the amp.

Best regards
Ben

 

[SimpleMind]

The change in noise frequency means you have went from half-wave rectification to full-wave as it should be. There is room for improvement in the B+ and bias supplies, based on what the schematic shows there very little actual filtering going on. Starting with the bias, you can insert a 47k resistor between the pots and ground to reduce the adjustment range so it won't be so touchy. If you can't get the bias up to +6 volts then this value can be reduced. In the bias supply itself, replace C402/403/404 all with 100uF 63V,remember to observe polarity, insert 100ohm resistors between C402-403 and between C403-404. This should help clean up the bias supply a fair amount. The B+ you can add a 10 watt 10-22ohm resistor after C401 then another 330-470uF cap after that. Larger values of resistor will give more ripple reduction but trades more B+ voltage reducing power output. This may be fine though since the difference in volume between 25 and 30 watts is very little. Also not thrilled about the way the fans are powered, using the negative regulator for the preamp section, a total bean counter move. Perhaps adding a provision to power them from an external 12V wall wart would be better.

HI Jeremy

I want to report on the progress I've made in the meain time - or lack off...

• I changed the capacitor values as suggested, cut the tracks and inserted 100 Ohm resistors instead -> this did nothing to eliminate the ripple, but it completely cut away the switch-off pop - a very nice side effect.
• I had to wait some days, until I got the 10 W resistor to add a second cap to C401. I simply put the resistor and the second cap (390µF/450V) in parallel to C401, which would be a somewhat big dimensioned snubber. But that changed nothing at all.

-- I'll load up an updated schematics, as soon as I have figured out how. --

May be I misunderstood your notion "The B+ you can add a 10 watt 10-22ohm resistor after C401 then another 330-470uF cap after that. " Schall I put the second cap + resistor in series with the first?. It is quite hard to fit the bulky cap anywhere on the power supply board without getting excessive lead lengthes. Currently I need about 80mm of wire, which might be detrimental?

During my work I also killed the preamp board again and now (after replacing Q315 and Q316 (J105 JFets) for a second time, the power amp section runs again, but the preamp gives out a much too low audio signal, which is heavily distoprted. So, I guess, I have killed another transistor... But that is secondary too me, as the power amp section works just as good as before, including the ripple noise.

Any help would be appreciated. Thanks again,

Ben

 

[Triang]

Dear all
I think in my AMC 3030 is broken volume/balance control, somebody knows what is this type, and where i can buy it?