diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Solid State (http://www.diyaudio.com/forums/solid-state/)
-   -   Help with ripple (http://www.diyaudio.com/forums/solid-state/150052-help-ripple.html)

akis 27th August 2009 08:55 PM

Help with ripple
 
3 Attachment(s)
I have been busy building the guitar amp which I have mentioned on these forums before.

I have a problem with hum on some of the preamplifier circuits. The background:

pre amp transformer: 9V AC, 50 VA, 230V (on 240 supply), 15% regulation (from tables).
Current drawn: 37mA.
Filter caps: 4000uF
Voltage DC: 15 V DC
Ripple current theoretical: 90mV p2p
Ripple current as seen on scope : +/-40mV (totally agrees with theoretical)
Ripple current on scope with additional 4700uF: +/- 24mV

From the above it can be seen that all is normal and as expected.

Except I did all my designs using my bench power supply, which has 0 ripple. When we introduce ripple into the equation bad things happen.

I have tried adding voltage regulation in the form of a LM317 and it did reduce the ripple to half, but did not eliminate it.

Therefore I now have a PCB which houses 4 pre-amplifiers (reverb out, reverb in, mixer and headphones), all built with discrete components, of which 2 are suffering badly with hum and the other 2 are immune. All pre-amps are bypassed locally with a 100uF and a 100nF ceramic.

So it happens both the hum-affected pre-amps are a new design of mine which I have not ever used before, works perfectly with the bench power supply but fails with mains. These pre-amps use a class B push-pull at the output (to be able to drive the headphones and the reverb springs).

I have two options:
1. eliminate ripple completely (as my bench power supply does) or
2. revisit the pre-amp designs that is suffering so badly.

I attach the problematic circuit and some pics of the half-finished project.

Please help if you can :-)

akis 27th August 2009 08:56 PM

2 Attachment(s)
This is the ripple as seen on a division of 20mV and a picture of the beautiful bench power supply ;)

akis 27th August 2009 09:10 PM

1 Attachment(s)
Ineterestingly this pre-amp with a huge 170x gain does not suffer from hum at all. I know the use of NPN followed by PNP amplifies hum out of phase and reduces it. Is this perhaps the way to go for the other circuit too (simply add a NPN stage at its front?)

wintermute 27th August 2009 09:26 PM

Hi Akis just a thought but have you tried moving the toroid further away? It may be airborne, and the new circuit may be more suceptable to this form of noise. I have that problem with the preamp in my integrated amp, picks up a lot of hum radiated by the torroid, move the torroid out of the case and it drops dramatically. I ended up putting a lead sheild over the torroid which improved things a lot.

Tony.

Greg Erskine 27th August 2009 09:28 PM

Don't be 100% focused on ripple, it might be a wiring/grounding issue or some other issue.

akis 27th August 2009 09:31 PM

Hi

I have also used an external transformer, about 3 feet away, with the two toroidals shown on the picture completely switched off. Same issue.

I have tried various combinations and the problem is caused by the ripple itself. My circuit is not designed to be immune to it.

I am in the process of adding an NPN transistor at the front with the same gain as the PNP that follows (x2), see if it will eliminate hum! Here's hoping.

fotios 27th August 2009 10:37 PM

Power supply ripple is eliminated by using some constant current sources in series with the supply rails. For example, you can connect a PNP transistor in common base mode in series with the positive rail comming from the smoothing caps. Connect the emitter as input and the collector as output. Connect the base via a 15-20K resistor to 0V rail. The inverse in the negative supply rail using a NPN transistor. Transistors must have a power rate according to the consumption of circuit feeded. This circuit can be used only in small power circuits. Not in power amplifiers!
Take a look in the site of Bryston which offers ready schematics of such circuits. Study the power supply schematics to get some ideas.
In power amplifiers, the ripple exists in supply rails regardless if we use monster smoothing caps. In the small signal stages of amplifier, where the problem of noise caused from ripple it is worse than the output stage, the use of constant current sources it is very common and eliminates completelly the ripple.
As for that said Greg Erskine, it is right by some way, but the existence of ripple worsen the problem of hum and buzz noise. Toroids, are :bs: by 99%. Those are apparently very well screened... ha, ha, ha! Those are in reality very nice transmission devices of strong electromagnetic flux in any direction!
You must be verry cautiouss as for their place and orientation against the pcbs of amplification. A general rule to avoid the infection of amplifiers from the straight magnetic flux transmited from the toroid, it is the place of xformer horizontal and the pcbs vertical (to form an "L" shape) or the inverse, and far away between them as possible.
The covers of the case, acts as well like mirrors and reflects in unpredictable directions the magnetic flux of toroid. In such a case, you might not have hum noise with the device oppened but when you mount the covers on the case, you might have noise! Amazing but real! Try some experiments on this.
I am tired from toroids. The best instrument to check the output noise it is a high sensitivity P.A. speaker (like a 103dB/1W) connected in output. Trust me, i have seen with my scope amplifiers with output noise of 2mVrms - which is theoretically inaudible - but the speaker connected in output reveals the very small - but audible! - buzz or hum noise exists yet.
I wish you success.

Fotios

kaos 27th August 2009 11:08 PM

If youíre sure itís the ripple thatís haunting you one simple possibility is to try a PI filter to reduce the ripple (possibly a triple stage). That technique has been used for a long time with designs like yours to keep power supply noise out of the works.

Wavebourn 27th August 2009 11:18 PM

Proper grounding/powering wires never hurt.

Think of ripple currents that charge caps, and of voltage drops on that wires.

akis 27th August 2009 11:35 PM

thanks for all comments.

The NPN transistor in front of the PNP did not fix it but resulted in a rather messier PCB.

However I eliminated ripple completely by using a typical common collector regulator comprising of a single NPN transistor in line with the positive rail (there is no negative rail), and a zener for voltage reference at its base bypased by a capacitor. Taking the voltage reference straight from the zener into the base of the transistor halves the ripple, but this is not good enough. So I used two RC networks in series before feeding the base and there is absolutely no ripple at all any more and the hum is gone!! I wonder why the LM317 regulator I tried earlier did not achieve the same (or better) result.


All times are GMT. The time now is 04:46 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2