Hi there,

I'm building a fairly simple switch box using some relays, but I've come across some trouble trying to filter out DC hum (the box is switching audio signals for guitar pedals, basically a whole bunch of ins/outs).

I've been trying to figure out what value capacitor/resistor would be necessary to create a low pass filter and kill the 50Hz hum. I'm working with the equation F=1/2piRC which I believe is correct for figuring this out.

My problem is that the total load on the power circuit will vary depending on how many relays are active, which as I understand (and I could be way off) would require the C value in the formula to vary as well.

Numbers are as follows: Freq. = 50Hz, voltage is 9VDC, total potential load is close to 4Kohm (each relay is 275ohm when on, plus wiring, switches etc). Ground loops aren't testable yet as I'm still designing the PCB trying to accomodate a filter, etc.

Any help here would be much appreciated, if I've asked an old question just point me in the right direction.

Thanks!

 

Your hum could be 50Hz or 100Hz. After the rectifier the frequency is 100 Hz. However trying to filter out the hum. Will also filter out the lower part of the audio spectrum. You switchbox what kind of supply voltage does it use. Can you for a short use batteries to power it. In order to see if the hum goes away?

 

If I run it on batteries there is no hum. The audio frequencies shouldn't be affected, as the power and audio lines are both separate. What I've read has lead me to believe I'd be putting the filter on the power line, not the audio.

 

Can you post some schematics. The problem could be in your power supply unit

 

The PSU is just a 'regulated' 9VDC power pack, I designed the unit so that it and all the effects pedals could be powered from the same source. As I said, I haven't got as far as building the unit yet, only prototyping some connections as tests.

In the test I have a relay with power connected to the coil, and the throws are routing a guitar signal to A or B. As soon as power is connected there's an awful lot of hum. There are no ground loops, nor do the power and audio signals ever come into contact, hence me turning to a LP filter.

 

But still we need some schematics here and perhaps some pictures also. And also elaborate on how power is distributed. I am not sure your assumptions are correct.

 

Ok, I dont have any software for drawing schematics, but here's a rough hand drawn diagram of how the test is wired.

The DC is running to the coil with a switch in between, and the audio is running to the common and throw terminals. The whole design is based on this, with the power lines for the coils linked in parallel and the audio signals chained in series.

A and B + both run to separate amp channels for the test, both produce hum when power is connected, and its not any different whether the coil is open or closed.

 

Just to be sure. The coils and the audio signal do not share any common ground?

 

No, they don't. The power is earthing via the -ve contact on the power pack and the audio earths via the input jacks on the amp.

Going full scale with a metal case I was planning on keeping the 2 separate also, using the case as the audio earth in order to shield the wiring inside.

 

Are your power supply close to relay. Some switchmode powers can be noisy. Try to move power supply away from the coils as a test. And be sure that audio lines also are are kept away from anything that have with to do with the power supply

 

The PSU is just a 'regulated' 9VDC power pack...

Just to be clear, do you mean you used a voltage regulator IC to filter out PSU ripple? Did you use the recommended caps on the in and out?

You said it works fine on battery. Is that still true if everything is held in similar placement, in other words to address the issue of nearby noise interference?

 

I'm assuming everything is correct with the PSU, it's just a store bought wall wart supply for guitar pedals. I've tried similar projects in the past, and my solution has always been to use battery power. This one is considerably bigger though, so batteries aren't reall an option. I'm about to shift the prototype to a board, just to rule out any possible induction (despite the shielded wires...).

I'll report back with photos once I get home and give it a go, thanks for all the help so far!

 

Ok, so I moved the setup to a board and still got the same results. The hum may have been a little less, but still very noticeable. Here's some photos of the board, if that helps anyone. The little bit of wire that looks like its bridging in the top left isn't, I moved it and tested again.

As I said, its a regulated DC wall wart providing power. I've used the supply with pedals etc. and its never been a problem. Obviously this is a different scenario, I'm just pointing out that its been fine in other applications.

 

... its a regulated DC wall wart ...

That's the problem. "Regulated wall wart" is an oxymoron; wall warts are not regulated and pass a lot of the ripple along. That ripple can cause hum in circuits not designed to reject it.

Depending on what the wall wart is really putting out (probably a lot more than 9V), it's possible you could add a 9v regulator IC that would reject most of the ripple. Regulators have to operate with a voltage drop (called the dropout), and you may not have enough headroom. So you may need a 12v wall wart.

 

Or it could be that the unterminated terminals are causing hum pickup.....
An O'Scope would help to isolate the offending signal source.
..
Here is a simplified - non technical troubleshooting suggestion:::::
Then again, you could always use a whole bunch of batteries and makeshift a DC battery for this device and see if the hum goes away. If it does - it is your Wal Wart. If it doesn't - you have other issues.
...
If it turns out to be the walWart - then just build a capacitive filter (the 50Hz reject filter may sound good, but will it actually work as well as you want it to? Or would an actual capacitive filter be better?
..
Dave
Phoenix, AZ

 

I haven't checked the output on the power pack, I'd assumed that it was basically correct as it works with everything else. I'll check that out when I get a chance tomorrow.

As for the capacitive filter, that was my initial problem - the load from the full circuit will vary considerably, so I was unsure how to build a filter that would accommodate it. Could I a) place an individual filter on each of the relays (they are powered in parallel) or b) build a filter that can deal with the varying load or would c) the regulator IC be the best option?

Thanks again

 

... would c) the regulator IC be the best option?

Cheap and almost 100% effective wins my vote.

 

Awesome, thanks wayneh.

Would something like this be suitable? http://www.jaycar.com.au/productView.asp?ID=ZV1640 (there's a .pdf in that link with more info, the descriptions from jaycar are pretty basic)

This may also be a dumb question, but I haven't dealt with voltage regulation too much - would running the regulator off a 12VDC/2.5A adaptor, with the output voltage set at 9V/3A be a viable option, or would it need something with a higher current output? Or would this require building a 12VDC (or more) PSU?

Essentially, would it be sensible to build the regulator into the switching unit so that it can be powered off a store bought PSU? Your suggestions have got me thinking that perhaps I can run this box on 15VDC/4A, and have 12VDC and 9VDC step down regulators in the box. I can then set up a series of 15, 12 and 9VDC outputs on the back of the box to power 10-15 pedals that require one of the three voltages.

Again, apologies if I've missed the mark entirely, I'm just trying to come up with the most efficient and compact way of condensing what has become a mess around my feet!

Thanks

 

That IC sounds interesting but I have no experience with it.

I do agree it sounds like time to sit back and think about the big picture. If you're going to put in the time to make something, it should be well planned. In my experience, you usually wish afterwards that you'd added more versatility or whatever. Initial plans tend to be narrow.

Anyway, think about how much power is needed at each voltage. Add in however much "headroom" you think you might ever want at each voltage. Start the design process from there and be sure to involve the pros here.

One think I'd consider in your situation is to start with a computer PSU or maybe a laptop charger brick, a switch-mode power supply or SMPS. These things are pretty cheap these days, efficient, small, rugged, self-protected and so on. You'll obviously need one rated at a power level at least 20% above your anticipated max continuous power level.

The details of stepping the voltages down from an SMPS, at the currents you need, is not something I can address. At less than 1A or so, I'd just use a 78XX linear voltage regulator. But the losses - and heat production - are just too much at higher currents.

 

Thanks for the advice.

I've come to the conclusion that I'll run it off a 20V/4.5A laptop brick (going for about $20AUD on eBay, doesn't specify regulation but is that such a problem now?), and have that power 4 regulators - a 15V, 12V and two 9V (one pedal requires an isolated 9V supply). I'll use the adjustable 3A regulators for each, as I doubt I'll ever pull 3A across all of them combined at any one time, let alone out of one individually.

http://docs-asia.electrocomponents.com/webdocs/06e8/0900766b806e8d55.pdf
That's the spec sheet for the regulators, I was planning on using the example circuit on p.7 which is the "adjustable regulator with improved ripple rejection".

These will feed about 20 DC sockets, each with a 3-way switch to select the output voltage. The relays and LEDs will be driven off the main 9V circuit with diodes tied across the coils to prevent any popping when they switch off.

The regulator boards will be separated from each other, and the main relay/audio board. As far as I understand, the regulators should eliminate any hum in the DC (and hence prevent the hum induced in the audio line) and the separation of the regulator boards is just a bit of insurance in my mind.

Does that sound feasible/sensible? I thought I'd put that forward before I get too far into the design.... Thanks again!