Problems Preventing hum in a small case

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I've just finished a new gainclone (Brians boards rev A, LM3875) in a small aluminium case( 50mm*160*110). Everthing is fitted in the case thightly, so including PSU, trany (120Va), potmeter, fuse, switch etc. The case is snugged up so thight that you might squeeze some beans in between but that is just about it. It is a neat litlle case, I'll post pictures when done. This is my third GC and 6th diy amp.

The amp has > 10.000uF, the originals Pana FC 1500 on the boards, which are snuberized a la Carlosfm and two 2200uf closer to the rectifiers.

Becaus I was already prepared for, well, at least challenges I wired the inputs to the potmeter and from the potmeter to the amplifier boards with tiny (2,5 mm) shielded coax cables.

Nevertheless I've got a big hum on the right channel and it does'nt seem to work proper either. There is no music, just hum. The left channel is working fine with almost no hum.

Grounding is by connecting the the channel ground of the boards to the case at one spot.

Inputs are insulated from the case and the channel shields ar tied to the pot. The input is not grounded, not before and not after the pot

The hum on the right channel is there, even if the potmeter (10K lin cheap pot) is closed. Again the left channel functions fine and reacts on the pot like it should.

I've taken everything out of the case and when I lift the coax substantially the hum diminishes until it's reasonable quit.

Apart of the hum I think the right channel might be oscilating because when it hums it also gets quite hot quickly, much quiker than the left channel.

I've checked all the wires three times at least and even replaced the chip.

-How can I check whether the channel is oscilating, unfortunately I have no scope.
- would shielded microphone cable (i.e. Prefer or alike) improve the situation on picking up the hum?

Are there any other suggestions that I might try to solve the hum and get the channel working properly.

Your help will be highly appreciated.


P.S. never use silicon heat conductive sheet that you have not tested before. The one that I had was sold as electrical insulating but the resistance was only 150 ohm which is not enough by far.
 
Here is a picture of the internals

gc_120_small.JPG


I first also had a problem with the aluminum bars that hold the chips which resulted in relatively high DC on the outputs. It was because of the silicon thermal conductive sheet. After removal the problem was gone and DC on the outputs is less than 30mV without a zobel

Could a zobel prevent possible oscilation?
 
I've checked my earthing, it's ok. Unless my chosen earthing layout is wrong (wire from the amp boards channel ground to the chassis, left and right connecting at the same chassis screw where the mains earthing for the case is fixed.)

I've noticed that some people use a 10 ohm resistor between the chanel grounds and the chassis. could that be a possible solution? would 0.6 w be ok or do I need a larger wattage?
 
Nordic,

Basically I have that erathing layout, the channel ground from the amplifier board is the same as the one that is shown between the caps.

I do not have a transformer with a center tap, just with dual secondary windings. The earthing layout worked like this in previous gainclones. That is why I'm suspecting something else

I think the 25V AC secondaries running close to the input and to the output might create the problem. With such a small case the layout possibilities are limited, specially when you want everything to fit in and have AC and DC.

I'm not giving up on the small case yet. I will try zobels and maybe have the input grounded at the RCA (through a resistor)

If that does'nt work I might use double shielded cables with the outer shield connected to ground on one side.
 
I posted that circuit to give you some guidance on the resistor you wanted to add... basically take the 2 biggest diodes you can find, reverse them, add 100nf cap and your 10R 2W resistor, all in parallel.....one end connects to PGS the other to the case, the mains earth then connecs to the case... make sure your input grounds are isolated from the case too... Right now my money is on hum picked up fromthe traffo... I think you mentioned one channel being worse... my money is on it being the one closest to the trafo... !

I used an even smaller box for dual monoblock, Unfortunately the Transformer needs to goo offboard.. actualy worked out quite well, as mine stands on the ground behind the hifi stand, out of sight.
 
Nordic,

Ah, I see what you ment. :smash: that I overlooked it before.

I have 6A diodes that will function nicely, the cap is also no problem but I cuurently have only 0.6w resistors in that value or very big 17w resistors in values that come close.

As space is at a premium here, is it possible to use a 0.6w as it is parallel to a 6A diode. I gues that in this case a small multilayer ceramic will do as well although I might have mkp filmcaps in that value.

I'll heat up the Iron now:D
 
Yeah, I have been using small MKP 100Vs but have since been told that it would be safer to use a mains voltage rated cap there for protection... You can test a bit with the small resistor, but I would replace it when you get around to it... I got some 2W resistors which are only a bit bigger than a 1W ceramic... cheap and nasty, but they do nothing, unless you have a short circuit anyway... so I just went with small fireproof ones...

As I said this story sits between PCU and case and between the amp's 0V and case you also put a 100n (0.1uf) cap and 100R , I believe it also reduces the likelyhood of your case acting like an arial...
 
Well,

I've solved one additional problem amd that is a broken coax. sometimes it conducts and sometimes it does'nt.

An aditional problem is that, although the hum is less, I only installed the resistor quickly, the channel is dead. I've already changed out the chip, there is power to the ampboard but it remains silent except for hum.

I've gotto solve this first before I can take a go at the humm
 
yeah, I have some ultra shielded cable, it consist of three concentric rings of braids of wire around a teflon core. the two inner rings are silver and the outer is copper. It's sooooo thick I can never get it in the case though.

I'll keep on trying to find a solution. I guess that I'm not the only one trying such a small case which contains both amp and PSU.

I know Peter Daniels has done some very nice cases although usually he keeps the PSU separate.

I'll see what I can come up with. Maybe the microphone cable is usefull because the shield is not used for the signal and I could erth the shield.

But first I will have to fix that one channel.
 
Been a while since I started this topic. I've finalised the amplifier (and did many other projects as well)

I had to fix a broken cable three times (never had that before), I changed out a chip because of short, I found a set of premium parts (rikkens ohms , caddocs, black gates etc.) so I changed them out. Changed the Speaker connectors for serious ones.

But most important of all I got rid of the hum. It turned out to be two little wires to power the led. Normally I power the led from additional windings on the transformer but this transformer is filled with moulded plastic. I tried with drilling a little hole but I needed to many windings to fit the hole.

I ended up with powering the led of the rail voltage with 70k resistence with two straight wires from the board alongside the tranny to the front.

The hum is almost gone since I twisted these wires , only a slight hiss is audible close to the amp/speakers.

I think this hiss is due to a very cheap radio ohm pot with a too low impedance which is the next (and last?) item to be replaced.

So: TWIST!!!, it does make a difference, especially when you are near the tranny and in very small cases.

edit: still to execute the double diode, film cap, resitor for grounding, at this moment still only a couple of resistors are used
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.