Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Tread power supply has high ripple
New Posts  All Forums:Forum Nav:

Tread power supply has high ripple

post #1 of 20
9/1/10 at 3:11am
Thread Starter 

I have just built a Tread power supply with one of Tangents complete kits. It works as expected with one exception, I get fairly high ripple on the DC out at about 45 mVrms. This is measured with a Uni-Trend UT60E which is not of Fluke quality, but a fair bit above any low budget DMMs. It meassures True RMS. So I don't think it is a measurement error. I tried too hook it up to a 24 V transformer as well as an unregulated 24 VAC walwart. Same problem...

 

I measured a random AC/DC walwart for reference and it measured at about 5-10 mVrms lower ripple! Now, something is clearly wrong with my build. Any ideas guys?

 

The tantalums are maybe a source of problem. I built this tread a while ago and tested them not before now, and I remember that the parts I got differed slightly from the scematics. But there was a large and a a small tantalum and I put the large one in C7 and the small one in C8. Did I get it the wrong way around?

 

Any other ideas of what causes this problem?


Edited by barsk - 9/1/10 at 3:31am
Reply
post #2 of 20
9/1/10 at 7:14am

Can you measure frequency?

Is the ripple 60 or 120hz? If its anything else something is oscillating.

 

Things can and do oscillate at random low frequencies if you are adequately unlucky.

Reply
post #3 of 20
9/1/10 at 7:20am
Thread Starter 

I did try to measure frequency and got 0 Hz. But I will try again to be sure.

Hmm, what in there is oscillating then? How do I track it down and fix it? I mean, a Tread does not have that much in it really...

Reply
post #4 of 20
9/1/10 at 8:07am

Try putting a capacitor in series with the meter leads... see what happens to your ripple measurement.

Reply
post #5 of 20
9/1/10 at 8:10am

How can you be absolutely sure that your meter will accurately measure ac riding on top of dc?

 

You should try to connect a cap to the output and measure using ac on you meter that way and compare the differences. That way you're measuring as only without the dc component.

 

EDIT:

What Pars said.... :-)

Reply
post #6 of 20
9/1/10 at 8:14am

The LM-317 is a deceptively complicated device. 

 

As another thought, how much voltage is across the LM-317? Vin(minus)Vout? The LM-317 does require a few volts across it to work at its best.

Reply
post #7 of 20
9/1/10 at 10:34am
Quote:
Originally Posted by barsk View Post

I remember that the parts I got differed slightly from the scematics.

 

18 months ago I updated the schematic to better match the kit contents.  Both schematics were correct, just differing slightly in part choices.  I decided it was best to go with the kit contents, since those who assembled their own would know why their part choices differed.

 

You're not building a kit older than 18 months, are you?
 

If not, which parts differed?

 

 

Quote:
But there was a large and a a small tantalum and I put the large one in C7 and the small one in C8. Did I get it the wrong way around?

 

You did right with the sizes.  But, a common mistake here is to get the polarity wrong.  Tantalums have their positive leg marked (matching a + on the TREAD PCB) instead of their negative as with electrolytics.  If you put them in backwards, they can fail without actually exploding, usually then causing the supply to misbehave.

Reply
post #8 of 20
9/1/10 at 12:15pm
Thread Starter 

@Nikongod: The transformer is giving me about 33V unregulated AC, and at the moment I am only pulling 18 V, so the regulator should have plenty of voltage. Did get a frequency measurement. 50 Hz it is, which corresponds with our 230 V/50 Hz in Sweden. Seems the filtering isn't doing its job?

 

@Pars and @digger945: Good suggestion. A bit embarrasing, that I did not conclude that myself ;). Anyway I get some confusing results. Depending on which capacitor I get somewhere between 5 and 100 mV. AC .Way above specs on 0.060 mVRms claimed by Tangent.

However, I also get 100 mV AC when the tread is switched off and only the wall plug is connected (I have a one-pole switch). So although the circuit is open I get some measurable 50 Hz AC voltage there... Now I am confused.

 

Edit: Changed word "circuit is broken" to "circuit is open". And filled in the correct unregulated AC power.


Edited by barsk - 9/1/10 at 1:18pm
Reply
post #9 of 20
9/1/10 at 1:39pm
Thread Starter 

Update. I hooked up the other Tread I had prebuilt. Tried it with both AC/AC walwart and transformer. Same results as with the other Tread. I am beginning to think this is a measurement problem and nothing else. Tried to measure with a another cap, I think a polyester film. And now I got more sensible readings, The readings slooowly approached zero. After a few minutes I was below 1 mVrms so I am leaning towards that there is no problem. I will try the same setup with the other Tread to confirm this.

Reply
post #10 of 20
9/1/10 at 9:35pm

Many cheaper meters have that problem.  They want you to measure AC *only* or DC *only*.  It's one of many reasons better meters are worth their price.

Reply
post #11 of 20
9/2/10 at 12:04am
Thread Starter 

A bit OT, but what to look for in the specs regarding this on a DMM. Or is it an "unspeced" property that one have to find out by trial and horror?

How are the Uni-trend DMMs regarded by the community? Spec wise they are in between the good (and expensive) and the bad (the chreap).

Mine, the UT60E is in the $60 range. 4000 count and about 0.5% error on Volt. It has a RS232 interface which you can use to display values on the computer and produce graphs, only 1 sec  poll interval though so not really usable as oscilloscope, TRMS, Auto range, Relative measuring. So it does a lot, and is easy to work with. Where do I need to go to better it?

Reply
post #12 of 20
9/2/10 at 4:38am

Look for AC-coupled measurements on the AC setting.

 

The Fluke 287 manual phrases it like so:

 

 

Quote:

AC mV, AC V, AC μA, AC mA, and AC A specifications are ac-coupled

 

The Agilent 33410A manual is more forthcoming:

 

 

Quote:

 

The multimeter's ac voltage and ac current functions measure the ac–coupled true RMS value. In this Agilent meter, the “heating value” of only the ac components of the input waveform are measured (dc is rejected).

 

If the meter only has a "volts" setting and auto-selects DC or AC for you, it'll be AC-coupled, but it won't do the right thing for a test like this, because the DC component is so much higher.

 

As for the rest, I answer that and more in my article How to Buy a Multimeter.

Reply
post #13 of 20
9/2/10 at 4:51am
Thread Starter 

I 've read that article. Good reading.

 

The meter has a button for switching between AC and DC and autorange from there. From the manual, there is no mention of ac-coupling as I can see.

Well, I have to live with the meter I have for now. I just bought it...

If the need arises, I guess I will have to take the plunge into the wallet and buy a proper meter.

Reply
post #14 of 20
9/5/10 at 2:02am

I wouldn't worry about upgrading your meter just yet. Just rig a simple AC pass (route current through cap, short to ground through large-value bias resistor) or AC block filter (short to ground through cap). It's not ideal but it should get the job done.

Reply
post #15 of 20
9/5/10 at 2:03am
Thread Starter 

One question regarding the Tread and shorts. What happens if you accidentally short the output? Will it fry the Tread or will the fuse on the input (500 mA) blow first, and protect the Tread?

Or do I need a fuse on the output as well?

 

I've built a standalone power supply with a transformer and fuse on input and was hoping this would suffice, but I need to get confirmation on this....

Reply
New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Tread power supply has high ripple