Welcome Guest! To enable all features please Login or Register.

Notification

Icon
Error

Login


2 Pages12>
Options
View
Go to last post Go to first unread
agent_cooper79  
#1 Posted : Tuesday, December 8, 2009 2:59:20 PM(UTC)
agent_cooper79

Rank: Member

Groups: Member
Joined: 7/21/2008(UTC)
Posts: 9
Location: Sweden

Hi,

I am experiencing som problems with my two single rail placid supplies. Both share the exact same problem: ca 0.25 V p-p (measured without load) high frequency, ca 6 MHz, oscillations. I have built them exactly according to spec with the parts supplied in the kits by twisted pear. The regulators work fine in all other aspects; current and DC voltage can be set without any problems. I have found that the DC voltage is constant up to the maximum load current while the HF AC component varies with load current and goes from about 0.25 V p-p without load to about half that voltage when driving a 100 Ohm load (ca 5.2 V rail voltage). Both frequency, amplitude and waveform is the same (or very similar) for both supplies given the same test conditions.

None of the four rails on my two Placid BP show any symtoms and are as quiet as can be using the exact same test setup, AC supply, scope, cables etc.

Can anyone help me where to start looking?

Best,

Magnus G.
Brian Donegan  
#2 Posted : Tuesday, December 8, 2009 3:48:48 PM(UTC)
Brian Donegan

Rank: Administration

Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 2,865
Location: Massachusetts, USA

Thanks: 2 times
Was thanked: 138 time(s) in 133 post(s)
Please describe the load you are driving. Do you see the oscillation with no load connected?
agent_cooper79  
#3 Posted : Tuesday, December 8, 2009 4:01:27 PM(UTC)
agent_cooper79

Rank: Member

Groups: Member
Joined: 7/21/2008(UTC)
Posts: 9
Location: Sweden

The loads I used are simple resistors. The oscillation has its highest amplitude with no load resistor connected: 0.25 V p-p. With a 100 Ohm resistor the oscillation amplitude is halved: ca 0.12 v p-p.

Cheers

Magnus
Russ White  
#4 Posted : Thursday, December 10, 2009 1:15:41 PM(UTC)
Russ White

Rank: Administration

Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN

Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
You could try increasing the comp cap. And/or possibly removing the 1nf cap that forms the secondary feedback. I would try increasing the comp cap first.

I have personally never encountered this.

Edited by user Thursday, December 10, 2009 1:17:02 PM(UTC)  | Reason: Not specified

Russ White  
#5 Posted : Thursday, December 10, 2009 1:18:40 PM(UTC)
Russ White

Rank: Administration

Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN

Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
You could also reducde the resistor in the cfp. Say 221R.
agent_cooper79  
#6 Posted : Friday, December 11, 2009 4:07:01 AM(UTC)
agent_cooper79

Rank: Member

Groups: Member
Joined: 7/21/2008(UTC)
Posts: 9
Location: Sweden

The strangest part of all this is that both Single Placids behave exactly the same, while both bipolar Placids work perfectly.

I will try increasing the copensation cap (the 22 pF cap, right?) first and report back, thanks for the help.

/Magnus
Russ White  
#7 Posted : Monday, December 14, 2009 5:55:04 PM(UTC)
Russ White

Rank: Administration

Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN

Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
Yes, change the 22pf to something like 100pf and see if that sorts things.
agent_cooper79  
#8 Posted : Wednesday, January 27, 2010 8:20:40 AM(UTC)
agent_cooper79

Rank: Member

Groups: Member
Joined: 7/21/2008(UTC)
Posts: 9
Location: Sweden

Hi again Russ & Brian,

I have now had some time for further test and diagnosis of the two malfunctioning Single Rail Placids.

I have implemented the 100 pF compensation cap as proposed*, and the problems were indeed reduced. On one of the Placids it cured the problem completely, on the other it greatly damped the oscillation. Also, if one connects a load (1 k Ohm resistor) to the output of the still oscillating Placid, the remaining oscillation go away completely.

What I would like to know is:

--- Do you have an idea why the oscillation depends so much on load?

--- Can I be certain that it remains stable under dynamic load conditions, ie in use deiving a Buffalo / MUX?

Regards,

Magnus G.

*Actually, in order to reach the above described level of performance I had to parallell a 100 pF ceramic with the 20 pF ceramic included in the kit.
Russ White  
#9 Posted : Wednesday, January 27, 2010 5:21:01 PM(UTC)
Russ White

Rank: Administration

Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN

Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
Hi Magnus,

I have never observed this nor can I reproduce it, so I really am not sure why its happening for you. I do believe it is though and I am determined to help get to the bottom of it. :) We might just include a higher value in the kits from now on just to be on the safe side, but it is also possible it may work better with no comp cap at all.

I actually use no compensation cap when using the opa227 because it has very good internal compensation.

Did you try removing the 1nf cap? That cap and its series resistor are completely optional, and in fact possibly be the cause of what you are seeing. I generally don't use it. If the problem subsides after removing that cap I will amend the manual to make its use more clear.

Cheers!
Russ

Edited by user Wednesday, January 27, 2010 5:57:58 PM(UTC)  | Reason: Not specified

agent_cooper79  
#10 Posted : Thursday, January 28, 2010 7:25:12 AM(UTC)
agent_cooper79

Rank: Member

Groups: Member
Joined: 7/21/2008(UTC)
Posts: 9
Location: Sweden

Hi Russ,

Thanks for the fast reply! I will try experimenting with removing* the 1nF cap and also take the Placid to another workshop where I have proper tools and probes so I can scope different parts of the circuit and not just the outputs.

Will return with results as they become available (not until next week).

*Edited message

Edited by user Monday, February 15, 2010 4:59:19 AM(UTC)  | Reason: Not specified

Russ White  
#11 Posted : Thursday, January 28, 2010 7:49:10 AM(UTC)
Russ White

Rank: Administration

Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN

Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
Don't jumper the cap. That would be bad. You have to make it open. :)
jt211  
#12 Posted : Tuesday, February 9, 2010 8:14:32 AM(UTC)
jt211

Rank: Member

Groups: Member
Joined: 2/9/2010(UTC)
Posts: 2
Location: Poland

Russ,
I have buy two single and one dual Placids.
After I conncted Anagram DIY DAC I also get this oscilation
Whith load modified 3.3V / 190mA I get 30mV p-p HF noise
5V /130mA - 10mV
+12V / 200mA - 200mV
-12V / 200mA - 100mV
This HF noises has made my DAC crazy after few minutes.
This looks like pushed me to make new power supplies.
I install higher C bt this not help at all.
How I can fix this ???
regards
Russ White  
#13 Posted : Tuesday, February 9, 2010 12:52:51 PM(UTC)
Russ White

Rank: Administration

Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN

Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
I have never had any oscillation from the cct, but I may not have it configured the same as you.

Try removing the 1nf C.

Also make sure that you are shunting at least 40-50ma more then your load.

Edited by user Tuesday, February 9, 2010 12:57:58 PM(UTC)  | Reason: Not specified

jt211  
#14 Posted : Saturday, February 13, 2010 1:13:30 PM(UTC)
jt211

Rank: Member

Groups: Member
Joined: 2/9/2010(UTC)
Posts: 2
Location: Poland


Russ,
We remowe this 1nF C from +/- 12V which oscilated must and now -12V not work at all.
I give-up on this supplies and will make a new ones.
All 3 oscilated and this have effect on my DAC board.
DAC is a too expensive to made any more try with this supplies !!!
We not make any fouls on soldering as 3 guy check this.

Russ White wrote:
I have never had any oscillation from the cct, but I may not have it configured the same as you.

Try removing the 1nf C.

Also make sure that you are shunting at least 40-50ma more then your load.
Russ White  
#15 Posted : Saturday, February 13, 2010 1:17:20 PM(UTC)
Russ White

Rank: Administration

Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN

Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
Removing the cap should not have any such effect. Any chance you can post pictures?

I can't seem to make one oscillate, and I am just trying to figure out what could be going on with yours.

Sorry for your problems. Lets see if we can resolve them.

Cheers!
Russ
Russ White  
#16 Posted : Saturday, February 13, 2010 8:16:38 PM(UTC)
Russ White

Rank: Administration

Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN

Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
I have been working a while now to try to discover what might be going on.

It appears to me that sometimes it may indeed be better to use 150-220pf at the comp cap positions (C3/C10 on placid BP PCB version 2.1.1)and optionally you can leave C4/R7 and C8/R13 open.

We will make this kit change.

In the single rail placid (PCB version 2.0) C3 would be 150-220pf and optionally omit C4 and R7

I hope that helps, and sorry for not finding this earlier. I have just replaced my scope, and that helped quite a lot identifying this potential issue.

Now bear in mind, not everyone will see this potential issue. I have only one module on which I was able to get this to happen.

Cheers!
Russ
Russ White  
#17 Posted : Saturday, February 13, 2010 10:09:08 PM(UTC)
Russ White

Rank: Administration

Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN

Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
Also keep in mind the placid is tuned to have some decoupling capacitance at the load. So when you test it use something like 100uf.
agent_cooper79  
#18 Posted : Monday, February 15, 2010 12:28:54 AM(UTC)
agent_cooper79

Rank: Member

Groups: Member
Joined: 7/21/2008(UTC)
Posts: 9
Location: Sweden

Hi Russ,

I have bad news and good news... bad news first - right?

It turns out we had oscillations on ALL our supplies, ie two placid BPs and two single Placids were all oscillating. The reason we didn't discover that the Bipolars were oscillating the first time around was that the oscillations were slightly different in shape and therefor harder to spot on our old analoge 'scope (a straight line was showing, but it had a faint Corona from the oscillations that we missed the first time) d'oh! .

Now for the good news; :d/ For me and my friend the problems with oscillations were completely solved for all supplies by removing C4 (single rail) and C4, C8 (bipolar Placid). Regardless of loading conditions etc we did not see any oscillation with the earlier mentioned cap/s removed. We also went back to stock configuration for C3 without inducing any oscillation. A smaller cap in C3 seemed to help transient response (reduce output impedance) slightly.

Seeing how all our supplies were oscillating before removing C4 (single rail Placid) or C4, C8 (bipolar Placid) I would be very intersted to hear from other peoples experiences of PLacids regarding this problem. Think

Cheers

Magnus

Edited by user Monday, February 15, 2010 12:30:51 AM(UTC)  | Reason: Not specified

Russ White  
#19 Posted : Monday, February 15, 2010 5:01:09 AM(UTC)
Russ White

Rank: Administration

Groups: Administration, Customer
Joined: 10/24/2006(UTC)
Posts: 3,979
Location: Nashville, TN

Thanks: 25 times
Was thanked: 89 time(s) in 83 post(s)
Thanks Magnus,

Your result is as I suspected.

I added the R/C combo C4/R7(and its partners) too try to tune the output impedance. In the end I ended up removing them from the latest boards.

Thanks for helping out with this new circuit, and I apologize for not running into this earlier.

Here are some other things I found can help:

1) Don't use C4/R7(Plaicd 2.0/Placid BP 2.1.1) and C8/R13 on placid BP(2.1.1)
2) Increasing C3/C10(comp caps) while it does increase output impedance at very high frequency does make the supply more robust. Up to 220pf should still yield excellent results.
3) Decreasing the values of R5/R11(normally 221R) to about 75R also increases the phase margin.
4) Unless you wiring is very short, leave decoupling caps in supplied circuits! :)

Cheers!
Russ

Edited by user Monday, February 15, 2010 5:02:47 AM(UTC)  | Reason: Not specified

Beefy  
#20 Posted : Monday, February 15, 2010 7:18:46 AM(UTC)
Beefy

Rank: Advanced Member

Groups: Member
Joined: 5/9/2008(UTC)
Posts: 62

So what is the recommendation for those of us without scopes?
Rss Feed  Atom Feed
Users browsing this topic
2 Pages12>
Forum Jump  
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.