Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

[everell]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

You have blue wire going to 9.1k to ground. Remove the 9.1k resistor and replace it with a small piece of wire. This will short the blue wire to ground. Now measure the emitter of the optoisolator (pin 3) with respect to ground....should measure zero ohms. Now measure the feedback pin of the VIPer 27 with respect to optoisolator collector (pin 4)......should also measure zero ohms.

The VIPer 27 does not use that 9.1k resistor. Wiring to the optoisolator pins 3 & 4 is backwards from the VIPer 22A.

[ben7]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

Hmm... looks like the feedback loop is not working since it hit 10V.
How many volts at the feedback pin (careful, don't probe short!) (or measure across the transistor portion of the optoisolator)?

Since the VIPer27 is a pull down feedback, the optoisolator should be pulling the feedback pin down to 0.3V (saturation voltage) when the voltage is that high. See if the "output" side of the optoisolator is indeed saturated. If it is, then something's wrong with the VIPer27 circuit, else something's wrong with the sense circuit on the secondary...

I say the same. Also, there should be a small capacitor (10nF?) across the opto transistor, to help make it stable.

[eccerr0r]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

Hmm... looks like the feedback loop is not working since it hit 10V.
How many volts at the feedback pin (careful, don't probe short!) (or measure across the transistor portion of the optoisolator)?

Since the VIPer27 is a pull down feedback, the optoisolator should be pulling the feedback pin down to 0.3V (saturation voltage) when the voltage is that high. See if the "output" side of the optoisolator is indeed saturated. If it is, then something's wrong with the VIPer27 circuit, else something's wrong with the sense circuit on the secondary...

[mockingbird]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

Nice job, looks great.

[LDSisHere]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

I received my SOIC-8 Viper22A chips tonight. Do to a shipping error I did not get the switches I ordered, for the load tester. I was able to populate both the 22A board and the 27 board and see if they would power on. The 22A works and is stable. The 27 started off at over 10V then it got unstable with fluctuating voltages and shutting down. It was late so I did not go any farther.

[LDSisHere]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

I have some Viper22A SOIC-8 chips and the components to make a 5VSB load tester like the one Everell made. This way it should be easy to do side by side comparisons of the surface mount versions of the 27 and 22A.

I left the word "ordered" out of my previous statement. I will hopefully have them tomorrow if the USPS does what it us supposed to do.

I have finished the layout for the Viper22A SOIC-8. I have also etched both the board for the 22A and the revised 27 board. I have the parts to populate the 27 board so I plan on getting started on it ASAP. I will post updates on this project as soon as I have anything interesting to add.

Thanks for the help and support guys.

[Per Hansson]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

Thanks for all the help in trying to revive this supply and I apologize for being the cause of the problem.

Don't apologize for learning, while I'm mostly lurking in these 5VSB mod threads I find the whole procedure very interesting & rewarding

[ben7]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

Thanks for all the help in trying to revive this supply and I apologize for being the cause of the problem.

No need to apologize!

This is called learning!!!

I also hope to learn about what findings you have, from the two different SMPS chips, and their DIP/SOIC parts.

[retiredcaps]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

Thanks for all the help in trying to revive this supply and I apologize for being the cause of the problem.

On the upside, you learned a lot and may have a touch hold enabled Fluke.

[LDSisHere]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

I was rushing last night so the board is not some of my best work but it should be functional.

Good news and bad news.

Good news: The power supply is now working.

Bad news: I am an idiot.


In my haste to assemble the board last night I reversed the position of the resistor and diode.

Ben7, thanks to your advice I now have a much greater understanding of how the 5VSB circuit is wired and driven. I checked and rechecked the power supply components and PCB and I could not find any reason as to why it should not work. I figured either the optoisolator I used or one of the components on the Viper board was bad. It was when I started populating another Viper board to test that I realized the source of my problem. After putting the components in the correct location the supply started working properly.

I have some Viper22A SOIC-8 chips and the components to make a 5VSB load tester like the one Everell made. This way it should be easy to do side by side comparisons of the surface mount versions of the 27 and 22A.

Thanks for all the help in trying to revive this supply and I apologize for being the cause of the problem.

[ben7]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

(With the PSU off) Measure the resistance of all the pins on the primary side of the 5vsb transformer. You should find two windings. If you only find one, and you have checked all other possible combinations, then one of the windings is open circuit.

[retiredcaps]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

In the third picture I have numbered the points like was described in the links and circled where the zero ohm resistor is supposed to be moved. Does this look correct?

Yes, that is correct.

Once you move the 0 ohm resistor, in order to enable autohold, you have to turn off the multimeter and then turn it on while holding down the grey "range" button for at least 2 seconds.

If touch hold is enabled correctly, it will show no numerical reading on the display since it has not acquired a stable reading yet. Once you get a stable reading it should then beep and freeze the reading.

PS. In order to remove the pcb board from the case, you have to remove the fuses since there is a screw underneath that is attached to the case. Also try not to touch or move that black screw pot (that is for calibrating DC V).

[LDSisHere]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

I cannot find anything damaged on the power supply PCB. Other than the optoisolator which I replaced there was no visible damage to the components on the supply. The Viper board is new along with all the components on it. I have enough parts to build one more Viper board so I may do that and see what happens.

In case anyone is wondering I checked the Viper board for shorts before I used it. The trace that runs between the capacitor legs is not touching anything is should not be touching. I was rushing last night so the board is not some of my best work but it should be functional.

[eccerr0r]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

I have a feeling your main switching transistors (2x2SD4515) are probably still OK as they should not have taken the brunt of the discharge. I just hope the pcb wire tracks are not damaged from the main filter caps to the viper22a board (and anywhere else the +320 could have gone with the probe slip).

I'm thinking the VIPer board is not oscillating anymore and the 5VSB transformer is thus not being powered. With that dead, no power will get to the KA7500C. The VIPer should be pulling on the 5VSB transformer directly (I don't have the schematics or data sheets handy right now...)

[LDSisHere]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

It looks like the Fluke 21 and 75 are the same. If yes, then

https://groups.google.com/forum/?fro...cs/rgEo-5NkMOo

applies.

A picture of the Fluke 75 pcb with that resistor on pads 4 and 5.

http://files.radioscanner.ru/uploader/2011/fluke_75.jpg

I have attached pictures of my meter. It does look the same as the one you posted except mine does not have the top left lead connector. It also has Fluke 21 etched into the PCB.

In the third picture I have numbered the points like was described in the links and circled where the zero ohm resistor is supposed to be moved. Does this look correct?

Check fuse for zero ohms. Check viper board with dc voltmeter, plus probe on drain heat sink (pins 5,6,7,8) and negative probe on ground heat sink (pins 1 and 2) should be 320 volts. Check that you did not install the optoisolator backwards. Check main switching transistors for shorts. Then check the 7500.....check for 5 volts on Ref pin. Then tell us results on these checks.

Gnd -> Drain 327VDC
Optoisolator is orentated correctly.
main switching transistors - In the fourth picture (taken before the Viper mod) are the two transistors on the right the ones that you are refering too? If so the do not show as shorted.

On viper board:
Green wire (1) GND
Blue wire (2) 0V
Yellow wire (3) 0V
Brown wire (4) .8V

Viper22A VDD 7.53 (Internally generated I am assuming.)

The KA7500C has 0V VDD (pin 12) and 0V on the reference pin 14.

Could the transistors be bad without being shorted?

Like I said before I am at a loss at what I should do next so any help would be appreciated.

[everell]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

Check fuse for zero ohms. Check viper board with dc voltmeter, plus probe on drain heat sink (pins 5,6,7,8) and negative probe on ground heat sink (pins 1 and 2) should be 320 volts. Check that you did not install the optoisolator backwards. Check main switching transistors for shorts. Then check the 7500.....check for 5 volts on Ref pin. Then tell us results on these checks.

Having done the same, reminds me of the old proverb.....one awe shit wipes out all your atta boys! Indeed it does slow up progress. On the other hand, we work on these old power supplies because we want to learn something, sometimes not what we were expecting to learn.

[LDSisHere]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

Thank you very much, that information is very helpful. I saw your load tester earlier today when I was looking through old threads for more information. I intend to add the resistors to my Viper order.

I just finished building a new Viper22A board and replacing the blown optoisolator. The 5VSB did not come alive, not that I am surprised. Any ideas about what I should check next would be appreciated. Is there a chance that I burned up a transformer? This setback is really holding up my progress, I guess I should just look at it as another learning opportunity.

Thanks,
Lloyd

[everell]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

To better understand why some pwm chips worked well on 5vsb and why others didn't, I built a 5vsb load tester as described in this thread:

https://www.badcaps.net/forum/showth...hlight=everell

The idea is that by adding ten ohm resistors in parallel with the 5vsb circuit, I could approximate the loading of .5 amp, 1 amp, 1.5 amp, and 2 amp. Even more if needed. What I was looking for was how well the 5vsb was regulated. What I found was that sometimes the 5vsb was too low with no load, and sometimes the 5vsb went to almost zero when 3 or 4 resistors were added. Why?

The Vdd voltage is a determining factor. If the Vdd voltage from the feedback winding is too low, the pwm chip will be constantly trying to use the startup feature, but cannot produce full power while in the startup sequence. On the other end of the spectrum, as more resistors are added, more power is delivered to the regulated 5vsb, but the voltage on the feedback winding is NOT regulated. If the Vdd without added load resistors is high enough for proper operation, the circuit will output the 5 volts properly. But as more and more resistors are added, the Vdd increases. If it increases beyond the Vdd cutoff point, the pwm chip shuts down. So lets begin by looking at the DM311 chip.

The minimum Vdd operating voltage (from the feedback winding) is 9 volts. If it drops below 7 volts, it tries to restart. On the other end, if Vdd goes above 20 volts it goes into overvoltage shutdown. Using the Antec SP450 as an example:
No resistor 5vsb=5.08 volts Vdd=16.02 volts
one resistor 5vsb=3.81 volts Vdd=15.45 volts
two resistors 5vsb=3.38 volts Vdd=3.01 volts
three resistors 5vsb=2.47 volts Vdd=19 to 23 volts fluctuating
four resistors 5vsb=2.02 volts Vdd=12.1 volts

To better understand what was happening, now look at the measurements using the VIPer 22A. Minimum Vdd operating voltage is 14.5 volts, Vdd maximum voltage (overvoltage circuit kicks in) is 42 volts.
No resistor 5vsb=5.08 volts Vdd=14.8 volts
one resistor 5vsb=5.04 volts Vdd=20.0 volts
two resistors 5vsb=5.01 volts Vdd=24.0 volts
three resistors 5vsb=4.98 volts Vdd=26.7 volts
four resistors 5vsb=4.94 volts Vdd=29.6 volts

As can be seen, the feedback winding provided the minimum needed voltage with no added resistors. It was below the 42 volt maximum with four resistors. One or two more resistors could be added before the pwm chip would go into overvoltage shutdown. The result is that the 5vsb output values are consistent with what is expected.

Now you can see what happened with the DM311. When the first resistor was added, the Vdd was right at the cutoff point. When the second resistor was added, the DM311 was trying to shut down, then restart, then shut down, etc. Adding more resistors only made matters worse.

By making these comparisons on several power supplies with different pwm chip circuits, I conclude that the 5vsb transformer determines which chip you need to use. If you want to pick a chip that doesn't match the transformer parameters, then it will be necessary to construct a different 5vsb transformer. Then you are moving from hobbyist status to engineer status. That means time, money, and more patience than what I have to put into it at this point.

Based on my Vdd readings on a Antec SP450, when you add the second resistor, the Vdd has gone up to 24 volts. The Vdd max for the VIPer 25, 26, and 27 is 23 volts. I predict that the 5vsb will NOT work properly with more than two resistors which is a 1 amp load. So this series of VIPer would not be useful if you expect the power supply to deliver its full 2 amp on the 5vsb rail.

If you use the surface mount VIPer 22a and a large heat sink, you might get close to a two amp load. Just a guess.

[LDSisHere]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

It looks like the Fluke 21 and 75 are the same. If yes, then

https://groups.google.com/forum/?fro...cs/rgEo-5NkMOo

applies.

A picture of the Fluke 75 pcb with that resistor on pads 4 and 5.

http://files.radioscanner.ru/uploader/2011/fluke_75.jpg

I will have to open my meter up when I get home and see how it compares. From memory it does look the same but I have only had to change the battery a handful of times in the last 20 years so it will be interesting to see if they are the same.


I have attached the modified Viper27 schematic and PCB. I was able to change the pinout of the connector to match the one I used on the 22A so I will be able to plug this in without the need for an adapter. I am in the process of working up a board for the SOIC 22A and I hopefully will be able to post it tonight or tomorrow. Of course I will not be able to test the SOIC 22A until I get the chips but I will see if I can get them by Wednesday.

[retiredcaps]

Re: Antec Smart Power SP-350 5VSB mod using VIPer 22A pwm chip

If you have the original Fluke 21, it may be moddable to support touch hold by moving one resistor on the pcb. Let me do some research on this.

It looks like the Fluke 21 and 75 are the same. If yes, then

https://groups.google.com/forum/?fro...cs/rgEo-5NkMOo

applies.

A picture of the Fluke 75 pcb with that resistor on pads 4 and 5.

http://files.radioscanner.ru/uploader/2011/fluke_75.jpg