Setting up a ESP active crossover with gainclone

[Dan_Steele]

Right, Time for an update:

Since last time, I have removed the 47nf across the output RCA's of the x-over and replaced them with 5nF as per the data-sheet.

Also, I have added 220pF across the supply pins of all the op-amps in the X-over.

The effects of this have been that the level of noise has been reduced dramatically and turining the pot no longer induces a radio tuning like sound to be produced.

Also, increasing the output capacitance to .1uf now yields no improvement over 5nF, so I am confident I have go down the correct route here.

Now, the final noise appears to lie in the power supply cables that feed the amps.

Wiggling the cables around, moving them around changes the magnitude of the circa 50Hz hum (with a little bit of something else on top ~ no acces to scope at present)

The power spply cables are screened types, 4 core with shield.

Now I can only see this being down to 2 reasons;

1) The metal body of the power cable inlet (which isn't actually connected to anything ~ http://www.maplin.co.uk/Search.aspx?...source=15&SD=Y ) varies its connection of the chassis of the PSU which is connected to safety earty.

or

2) Changing the direction of the cable is like moving an antenna and is picking up a particular frequency. There are no flourescent lights nearby and I have tried turing off the kitchen refridgerator to no avail.

or 3) something really simple that I have overlooked.

I have tried a non-screened cable and the same thing happens (it occurs on both PSU's) and I have worked as an electrician in aerospace industry making such cables (but with a lot higher Quality Amphenol connectors) so I really hope its not my engineering at fault!!

I am going to try a 100uf bypassed with 100nf at the amp end between +ve and 0v and -ve and 0v in an attempt to reduced an interference picked up in the cable ~ any thoughts on this.

Regards

Dan

[Dan_Steele]

Right.. Another Update:

Couldn't find any 100uf's so I used 220uf Silmic 2's bypassed with 100nf between the -ve and gnd and +ve and gnd.

No change in hum.

So then I looked at it again, and decided to see what would happen if I plugged the left high frequency amp into 1 psu and the left low frequency into the other psu (the current set up has been 1 psu for each channel)

The hum vanished and all that was left was a faint hiss (which I decided I could live with)

So then I figured I had solved the problem in a round about way.

So preceeded to connect the right channel in the same fashion so I had one PSU for the high's and one for the low's.

Guess what, Hum returned!

So therefore as far as I can see im looking at some kind of loop caused by the common gnd in the ESP crossover and the PSU for the amps.

An obvious way (but not how Im considering doing it) would be to scrap the twin channel PSU and make 4 separate ones.

Regards

And as ever all opionions appreciated.

Dan

[pinkmouse]

OK, so you have a ground loop. Draw a picture of exactly how everything is connected and we should be able to help. In the meantime, are all your connectors insulated from the chassis?

[Dan_Steele]

Hi Pinkmouse,

I'll get on and do that,

All RCA's are insulated, the body of the IEC inlets are connected to the case (safety earth) and the body of the power supply umbilicals connectors are also connected but not connected to anything else.

In the X-over:

Body of ELNA switch is connected to chassis along with body of alps pot (but this chassis has no safety earth as power supply is separate ~ body of RCAs is connected to alps pot to reduce hum (but I guess this then means that the RCA's are connected to the chassis and therefore uninsulated.)

A drawing might make it look clearer

Regards

Dan

[Dan_Steele]

Right, I got a bit carried away and drew all of the interconecting wires etc. I do not have the circuit infront of because I'm at work at the moment, but im 99% that is how it is.

Only one channel is shown, but the problem happens when it is connected in the above fashion.

To re-instate, connect one amp to x-over, no hum, connect rca for second amp hum starts (works either way around i.e connecting hi then low or low then high)



Sorry if its a large file but I wanted all the details to be clear.

regards

Dan

[AndrewT]

Hi,
I got shot down for saying this in another thread, but I think I can duck quickly enough.

I believe the grounds for left and right should be kept separated all the way through the system, excepting ready built commercial gear which probably have a common ground for both channels.

I wonder also if that screen on the PSU cables which connects back to safety earth and also to audio ground in the crossover is partly the cause for the loop.

Can I ask you to go home and re-draw that diagram with ONLY the grounds and earths and zero volt lines and screens shown with all the component locations? I see no grounding on the majority of the RCAs.

Different issue and probably not related to your hum problem.
Are you sending unsmoothed DC down the PSU umbilicals to the power amps?

[Dan_Steele]

Cheers Andrew, well I can give it a try disconnecting them.

Only thing with the screens for left and right is that the x-over has a common gnd for all 4 outputs - leading back to the GND on the PSU board so without cutting tracks I can't see a simple way of solving this.

It seems intuative to separate the channels compltely, and was a bit dissapointed Rod hadn't implmented this in his design.

Edit: Well the only smoothing I can see could be down to the rather small 10uf's on Peter Daniels PSU boards, otherwise it is unsmoothed on the umbilicals - radiating noise?


Edit2: I will re-draw it over the weekend; the twin pairs from the RCA's are the centre and the outer screen. I didnt draw the RCA's for the right channel on the X-over. Missed them off on the input / selector as well.

Regards

Dan

[AndrewT]

the only smoothing I can see could be down to the rather small 10uf's on Peter Daniels PSU boards

Oh dear.
Peter D's chipamps are designed to suit his 96db/W speakers.

His philosophy appears to be to maximise sound quality in the mid and treble range, but now I'm putting words in his mouth.
I have argued with him about the lack of smoothing and it's effect on the amp's ability to play properly. We never reached an agreement, neither in maximum power delivery, nor in bass (or lack of it). I think the 10uFs are 1000uFs or 1500uFs but even this is far too low to fit with my understanding of how a power amplifier works.