Vout incorrect

First off, if you could send me your PI Expert or PI XLS design files, it would be really helpful. I've tried to recreate the design with what I can gather from your schematic but there area few things that are not totally clear from the schematic.

One big area that I'm seeing some differences in between my PI Expert design and yours is regarding the feedback components and phase boost network. Without the details of you're design though, I can't be sure this is the problem....

What do your load and line regulation curves look like? Have you tested your control loop yet? Are your efficiency numbers looking ok? Any other odd behavior?

If you could send over your PI Expert files along with your layout and transformer design, it would be really helpful.

Also, I noticed you had a note near the input of your schematic about the purpose of the 2 1M resisters. These appear to be Xcap discharge resisters. Regulatory agencies require that any residual charge left of on the Xcaps after disconnecting the power supply needs to discharge with a RC time constant of < 1sec.

So that I don't lose an opportunity to plug one of our other products, take a look at our CAPZero device. http://www.powerint.com/sites/default/files/product-docs/capzero_family_datasheet.pdf It eliminates the power dissipation associated with the Xcap discharge resisters while still meeting regulatory requirements for the RC discharge time. You might see a small bump in efficiency and a small drop in no-load input power by using something like CAPZero.

-The Traveler

Hi, Thanks for the reply and info. I'll certainly look at Capzero soon.
I've attached the PIExpert files as requested.
Cheers!

bbmilo -


OK...couple more requests and then I have a few observations and questions for you.


I've noticed some deviations between the PI Expert design and the PDF schematic you attached earlier. Since there are some differences, can you send me your transformer documentation. With this I can recreate your design as best as possible.


Also, a lot of power supply issues can come down to PCB layout. If you could send me a copy of your layout, that might be helpful as well. I don't need the actual layout files...a PDF would be fine as long as there arn't too many layers and I can see reference designators pretty well.


I was hoping you could give me some insight into the following deviations from the PI Expert design:
- I've noticed that your clamp design is fairly different than the one PI recommended. Have you verified that it's functioning correctly? Do you know what your clamp losses are?
- What are the specs on C3? (Your main output bulk cap). The voltage rating is a bit low and with the one cap you might run into some problems with ripple current ratings.
-There are several differences between the PI Expert control loop components and the ones in your schematic. Changing the components in PI Expert to reflect your schematic is giving me an estimated crossover point around 6-7kHz. This is WAY WAY too high. You should be aiming for around 1kHz. Have you verified the control loop measurements on the bench? If not, please make sure you do this...this is really important to the robustness of your power supply design.
- Just out of curiosity, why are you using two fuses in your design?


Sorry for all the questions, I hope they don't tie up too much of your time. If you can though, your transformer documentation and PCB layout would be helpful. If the schematic you attached earlier isn't up to date, please send a current one and I'll see if I can't help you find the source of your regulation issue. I have a feeling we'll need to change some of your feedback components. But again, I really need to see the transformer documentation and your PCB layout.

-The Traveler

Hi Traveler, thanks for the diagnostic questions ;-). The PCB layout is as attached.
1) Clamp design has been checked by my local Power Int engineer. No I have not verified them yet. We correspond through email and his response has been very slow.
2) C3 is rated 1500uF 25V 2.77A. I thought that should be enough for 12V output??
3) I now manage to get the ps to give the right Vout at no load by using this value for the control loop: R8 3.66k, R9 866R, C11 15nF, C12 68nF, R13 50k and R15 11k. With these values, the crossover point should be around 1kHz.
4) Two fuses are required for the safety standard we are designing for.

New problem:
PS enters auto-restart when even a very small load of <500mA is applied.
These are what I've attempted:
1) Changed D3 from the slow BAV103 to 1N4148.
2) Removed D4 to disable the snubber circuit.

Any idea what else could be the cause of the auto-restart?
Please help. Thanks.

Hi Traveler,
I've also increased the input capacitor, C9 to 133uF. There's a slight improvement. Now the PS can supply up to 0.7A at Vin of 95Vac then goes into auto-restart. Once that happens, it'll stay in hiccup mode even when the power supply is being recycled.
Any idea how can I fix this auto-restart problem?
Thanks.

bbmilo -


Can you send me your transformer documentation as well? That would really help me check/simulate your design. Any measurements you've done on your transformer would be helpful as well.


Do you have access to a gain/phase analyzer?

-The Traveler

Hi Traveler, the PIExpert uds file has already been attached in the original message at the top.

The latest update for this problem is that I found that Vb (bias voltage) goes up to 27V just before the unit enters auto-restart mode.

The power supply manage to supply full load when the V-pin is grounded. Does that mean that the transformer's leakage inductance is too high? How can I fix this problem?

Cheers,
bbmilo

bbmil0-


Sorry I missed the PI Expert file...I'm getting over a cold and pretty swamped today. It looks like we have two different forum threads going...


Given the output regulation and bias winding issues, I'd really recommend starting over with your transformer. It sounds like the regulation on the bias winding is so bad that it's affecting the regulation on your main output (since the control pin current is coming from the bias winding). Fiddling with feedback components won't address the issue if it's as fundamental as poor transformer design.


Depending on the number of prototype transformers you ordered and what you have in labstock, any chance you could hand wind a new transformer?


-The Traveler

Hi Traveler, since the problem has changed to auto-restart and not incorrect voltage, I thought I'd better start a new thread. Hmmmm we only ordered one sample but more samples will be coming soon from our manufacturer.

Actually I've been wanting to wind my own transformer. Can you please recommend some handy tools and commonly used parts/material that I should get to do this? Many thanks.

bbmilo -


Probably one of the most helpful things you'll need will be a coil winder. We currently use model 1200-1 from this company: http://www.adamsmaxwell.com/1250sacc.html


I'm sure there are other options out there...you don't need anything too fancy or too many bells and whistles. Even a non-motorized setup would be fine for small batches of prototypes. The one key feature that you would want though is a built-in method for counting turns.


Along with this, you'll need basic supplies: enameled magnet wire, TIW wire, some different thicknesses of copper foil (if you plan on doing foil windings) as well as various thicknesses and widths of tape.


Probably the biggest investment though will be the purchase of a low-frequency impedance and/or gain/phase analyzer. HP/Agilent had the 4194 which is still a great piece of equipment. Venable Instruments also makes some great equipment suited for this purpose. Ridley Engineering also has a great solution called the AP300.


The purpose of this equipment is so that you can make detailed measurements to tell you about your magnetics design. Often datasheet specs are wrong or misleading and it can be useful to have the ability to double-check the figures your magnetics vendor is giving you. Snagging a piece of equipment that can do both impedance *and* gain/phase analysis would be great as it would let you evaluate components as well as look at the control loops in your designs.


On a side note, check out Ridley Engineering's website. They have some great articles or power transformer design and how to accurately measure the characteristics of inductors and transformers. Very helpful.


On to your output regulation and bias voltage issue. I was picking the brain of another engineer in my group and said a simple thing to test would be to make sure your bias winding was wound correctly. His thought was that if it was strictly a coupling issue, you would tend to see the bias winding voltage increase along with increasing load on your secondary output. If it's a winding polarity issue, you're more likely to see the bias winding voltage increase with increasing ac input voltage.


This should be pretty easy to check either in circuit or out.
In circuit: power up the board at a fairly low input voltage (enough to get the controller switching). Using HV diff probes, look at your drain-source waveform as well as the secondary and bias winding voltage waveforms. If you have a clamp-on current probe, it could be helpful to look at your drain current as well. When your drain-source voltage is low (MOSFET is ON), your secondary and bias windings should see a negative voltage (their output diodes will be reversed biased). When the drain-source voltage is high (MOSFET is off) you should have a positive voltage on your secondary and bias windings (output diodes are conducting). If it turns out that your bias winding is 180deg out of phase, it could cause some of the problems you're seeing.


Out of circuit:
Using a function generator attached to the primary. Look at the resulting voltage waveforms on your bias and secondary windings. (Take care to mind the dot conventions). The bias and secondary should be 180deg out of phase with the primary voltage waveform.


Check this out and let me know what you find out.

-The Traveler

Hi Traveler, many thanks for all the recommendation and information on setting up facility for winding my own transformer. I'll certainly look at that soon.

Re the bias voltage issue, my bias winding voltage doesn't increase with ac input voltage but increase with load. Hence, the bias voltage goes over 27V and entering auto-restart mode as I increases the load. I have also checked the polarity and it is correct.

It was suggested in other forums to add a 33R resistor between D3 and C8 at the bias winding. It is said that the high bias voltage is caused by the high spike voltage of the bias winding due to high leakage inductance... which makes sense. So I did that and also had to reduce the optocoupler resistor on the feedback circuitry. Doing this has fixed the problem and the power supply is delivering full load at 14V. The bias voltage now starts at 7V at no load and 26V at full load.

Of course this is just a temporary solution. The next step I guess, is to make some transformers properly so to reduce the leakage inductance.

Thanks,
Juin