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

[ben7]

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

I am going to use what I have on hand. I plan on using a 9.1V and 12V in series so it should equate to ~21V.

I've had issues before, when trying to use two zeners in series.

Try to find a 16v or 18v or 20v zener. Usually, there will just be a number on the diode, like, '5.1' - that would be a 5.1v zener. Sometimes, they have numbers and letters, so you have to go digging on the web...

[budm]

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

That AUX winding is for supplying the power once the IC starts running, when the 5v supply output which is also used as the feedback to regulate the output on the FB pin of the IC, since AUX winding is on the same transformer, if the 5V output goes up the AUX voltage will go up at the same time, when the feedback from 5 volt regulate the Voltage both windings, the AUX out put will follow the 5V winding. The spike that damage the Viper is when the circuit first starts up, it can be higher than max VCC, that is why they (I) use ZENER diode to clamp down the AUX output just below the VCC. I used VIPER series our UPS design.
Use the scope and monitor the AUX winding when you first apply the power to the circuits, you will see what I am talking about.
By the way, I use 10 Ohms on the output of the AUX winding in my products. We use a lot of VIPER series in a lot of our products designs due to low cost and reliable as long as you watch for that spike during startup time due the lag time for the regulation loop to become stable.
Another improvement is to put 1/2W 6.2V Zener at the 5v output to clamp down the out to 6.2V max during start up or if the regulation loop has gone bad. Also to find out how well the regulation loop reaction time is, you can switch in and out the max load and min load while monitoring the AUX, and 5V winding outputs.
One good thing about buying 1000's of the IC a year is that you have access to the application engineers that will help you analyze and get the tips and tricks support.

[budm]

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

I subscribe to this magazine, it has lots of information which helps me a lot in designing the products. These are great website for learning about power electronics design:

http://powerelectronics.com/


By the way, if the drop in IC is not the same , the control loop circuits should also be adjusted.
http://www.edaboard.com/thread254203.html
http://www.beigebag.com/case_TL431.htm
http://www.sonsivri.to/forum/index.php?topic=23855.0
juts google control loop using TL431

[LDSisHere]

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

I finally got my 5VSB load tester built. I have been extremely busy so I have not had much time to spend on this project. I just had time to test it with the last Viper27 board I built to see if it worked properly. I have etched a Viper 27 board to use a Zener diode or two also but I did not have time to populate it tonight. I also just got in some 22V Zeners to use with the Viper27. In my quick test it would hold for just a few seconds with all the resistors on, which measured 2.7 ohms. After it started osculating on and off it would not stabilize until they all were turned back off. I hope to have some time tomorrow to work with it some more.

[LDSisHere]

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

I finally was able to get the new Viper27 board done. I used a 100 ohm resistor and a 22V zener to limit the Vdd. I think the 100 ohm resistor did enough that the 22V zener never came into play. The first five picture are of the Viper27 board in operation. You should be able to see the number of switches that were thrown in each picture. It was able to stay steady with up to three 10 ohm resistors in parallel providing the load. When the fourth one was thrown it would hold for just a short while then it would start resetting. The interesting part was that the Vdd never got higher than 18.1V. It may be that a lower value resistor could make the difference but I have not tested that theory.

The next six picture show the SOIC8 Viper22a board being tested. The interesting part of this test to me was that it acted much the same as the 27 the main difference was that the Vdd climbed much higher as would be expected. It climbed to 26.1V when all four resistors were engaged and it too started resetting. It would hold steady with three resistors like the 27.

I also did a little testing of the DIP Viper22a to see how it would perform and it would hold steady with all four resistors engaged.

I do not know if the SOIC8 Viper22a version can be modified where it would hold under a higher load as it may just be at its' design limits. I do think that more power may be squeezed out of the Viper27 but it too may have been running at its' operational limits. I think I will try a lower value of resistor to see what that does to the circuit when I get a chance.

[budm]

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

Three 10 Ohms in parallel (= 3.33 Ohms) to load the 5V, that is only about 1.5A. How much was the original 5v circuits supposes to be able to handle?

Viper22A
Table 1. Typical power capability
Mains type SO-8 DIP-8
European (195 - 265 Vac) 12 W 20 W
US / wide range (85 - 265 Vac) 7 W 12 W

SO-8 7WATTS, DIP8 12WATTS

So you are using SO-8
So 5V @1.5A = 7.5W if 100% efficiency. I do not think you will get reliable 1.5A from the setup.

If the transformer VA rating is not made to deliver more than 1.5A, then it will not matter much how hard you drive the transformer.

[everell]

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

With the 100 ohm resistor, what were the Vdd values for each of the ten ohm resistors? You would think that reducing the value of the resistor would help with all four resistors in. Actually, just the opposite is true, because when the chip is trying to shut down from overvoltage, it REDUCES the Vdd as it oscillates.

Write down the Vdd measurement for each of the first three resistors, then CALCULATE what the next value should be. The Vdd should increase for each resistor just like the VIPer 22A DIP circuit did.

[tom66]

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

A normal flyback is wound with the aux winding the same way as the main output winding. That way it is fixed at a constant ratio to the main winding.

But some self-oscillating circuits do it differently, which could be why you are seeing excessive aux outputs; you're picking up the "back emf" (I cannot recall the correct term.) This varies with load. What does reversing the aux winding do?

[LDSisHere]

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

Three 10 Ohms in parallel (= 3.33 Ohms) to load the 5V, that is only about 1.5A. How much was the original 5v circuits supposes to be able to handle?

Viper22A
Table 1. Typical power capability
Mains type SO-8 DIP-8
European (195 - 265 Vac) 12 W 20 W
US / wide range (85 - 265 Vac) 7 W 12 W

SO-8 7WATTS, DIP8 12WATTS

So you are using SO-8
So 5V @1.5A = 7.5W if 100% efficiency. I do not think you will get reliable 1.5A from the setup.

If the transformer VA rating is not made to deliver more than 1.5A, then it will not matter much how hard you drive the transformer.

This was discussed on the first page of this thread, it is the reason the Viper27 is being used as it should be able to match the output of the DIP version of the Viper22A.

Viper27: US / wide range (85 - 265 Vac) 12 W - Open frame

The SOIC-8 Viper22a version is only being used for comparison purposes. It was expected to be limited to 1.5A or less of output.

I still want to do a SO-8 version of this mod but when I was looking at the datasheet I noticed that it was good for 7W (US) where the dip version was good for 12W (US). Would the reduced power capability be an issue if the SO-8 version was used?

Well, most of the time, the standby power has to power the smps control chip too (ex, a TL494). Most PSUs have a 5vsb rating of 1.5 to 2A. So the 7W would only do about 1.4 amps. You could use the SMD version (16-pin SOIC) of the VIPer25

Maybe you should start by making a surfact mount board using the Viper 22A even if it has less power. Then develop a board using the Viper 27.

For the concerns about the transformer being able to deliver the 2.0A I am attaching a picture of the power supply panel information and my tests with the DIP version of the Viper22a also confirmed to me that it could deliver the required power.

I also did a little testing of the DIP Viper22a to see how it would perform and it would hold steady with all four resistors engaged.

The four resistors in parallel measured 2.7 ohms so at 5V that equals 1.85 Amps which is close enough to 2.0A to suit my purposes.

With the 100 ohm resistor, what were the Vdd values for each of the ten ohm resistors? You would think that reducing the value of the resistor would help with all four resistors in. Actually, just the opposite is true, because when the chip is trying to shut down from overvoltage, it REDUCES the Vdd as it oscillates.

Write down the Vdd measurement for each of the first three resistors, then CALCULATE what the next value should be. The Vdd should increase for each resistor just like the VIPer 22A DIP circuit did.

Everell I have attached an image of the schematic for the Viper27 circuit. Vdd is going through a 100 ohm 1/2W resistor and the Vdd also has a 22V zener attached to it but it should not have come into play according to my tests. If you look at picture 4 in post #85 you will see the Vdd is only 18.08V. In picture 5 you will see the same Vdd but the Viper starts to reset itself when attached to a 2.7 ohm load. In picture 1 you will see a picture of the Vdd at 10.90V when the Viper27 only has a 470 ohm resistor and LED attached to it.

The SOIC-8 Viper22a had a starting Vdd of 13.73V with only the 470 ohm resistor and LED attached to it. (Post #85 picture #6) The Vdd on this board peaked at 26.06V with a 2.7 ohm load. (Post #85 picture #10)

My initial thoughts on lowering the Vdd resistor value was that it was dropping Idd too much and causing it to shut down under high load, however I now do not believe that is the case since the Viper27's max Idd is 25mA. I may still try a lower value resistor just to see what happens. I do not think it would hurt anything and it would give us more information about what the circuit is actually doing and maybe give some clues as to why it is shutting down.

[everell]

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

Now that you have a load tester, try these tests:

Using VIPer 22a 8pin DIP circuit, run 5vsb for an hour with 4 resistors loading. Should do fine.

Using VIPer 22a SOIC chip circuit, run 5vsb for an hour with 3 resistors loading. Hopefully it will do fine.

Using VIPer 27 chip circuit, run 5vsb for an hour with 3 resistors loading. Maybe it will do fine.

Using VIPer 27 chip circuit, run 5vsb for a few minutes with 4 resistors AFTER reducing resistor to 75 or 82 ohms. Will it stabilize?

Using VIPer 27 chip circuit, run 5vsb for a few minutes with 4 resistors AFTER increasing resistor to 110 or 120 ohms. Will it stabilize?

Maybe BOTH soic chips just can't handle the power. Perhaps you should consider building a 8 pin DIP VIPer 27 circuit. Then test it to see if it can handle the 4 resistor loading.

When using a soic chip, we might have to use it to drive a power FET like in the Thermaltake Toughpower circuit I posted earlier.

[budm]

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

Viper22a: That is dip 8 can handle more power than the SO-8, four 10 Ohm resistors in parallel is 2A = 10Watts which Dip8 can handle.
What you see with SO-8 with four 10 Ohms is normal because you are trying to get more than it can handle.
Also what was original circuit output current rating, that will limit how much you can pull from the transformer, transformer can only handle so much per VA rating. You can try driving the transformer with higher rating SMPS IC, but if you do not increase the VA rating of the transformer, you will not get the out power that you want.

[LDSisHere]

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

Now that you have a load tester, try these tests:

Using VIPer 22a 8pin DIP circuit, run 5vsb for an hour with 4 resistors loading. Should do fine.

Using VIPer 22a SOIC chip circuit, run 5vsb for an hour with 3 resistors loading. Hopefully it will do fine.

Using VIPer 27 chip circuit, run 5vsb for an hour with 3 resistors loading. Maybe it will do fine.

Using VIPer 27 chip circuit, run 5vsb for a few minutes with 4 resistors AFTER reducing resistor to 75 or 82 ohms. Will it stabilize?

Using VIPer 27 chip circuit, run 5vsb for a few minutes with 4 resistors AFTER increasing resistor to 110 or 120 ohms. Will it stabilize?

Maybe BOTH soic chips just can't handle the power. Perhaps you should consider building a 8 pin DIP VIPer 27 circuit. Then test it to see if it can handle the 4 resistor loading.

When using a soic chip, we might have to use it to drive a power FET like in the Thermaltake Toughpower circuit I posted earlier.

I have a problem with the load tester that I made that will have to be corrected before I will be able to run it for an hour at a time. I think I made a bad choice in resistors. They were getting to almost 200 deg F after just ~10 minutes of running so I am going to have to get some different ones. This is a link to the ones I bought: http://www.digikey.com/product-detai...TCT-ND/3114511 I am now thinking I should have went with the "cement" type which were cheaper to anyway.

I will see if I can find your post about the Toughpower as the idea of driving a FET sounds intriguing and may be a much better idea for this project.

[everell]

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

For convenience, I'll just post it again here. To configure the pwm chip with a FET power driver, you should be able to use the power transistor in the original two transistor circuit. Just eliminate the small transistor (and its associated components) and use the pwm chip to drive it.

You might want to find a different power supply to do this experiment with - most any old crappy power supply should do fine.

[LDSisHere]

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

I have finally made some progress on testing the various VIPer chips and I thought I would share the results.

The big hold up for me was a load tester that I trusted to run for more than a few minutes at a time. The the resistors on my first one would get close to 200 Deg F in under 10 minutes so I was afraid to run it for a longer period of time. The first four pictures are of my new 5VSB load tester. I have five 10 ohm 10W resistors in parallel with each controlled with a switch. I mounted them an old P3 heat sink in a gutted ATX power supply. I left the fan which I power with a separate 12V supply so it has no bearing on the tests.

1 resistor = 10.1 ohm
2 resistor = 5.1 ohm
3 resistor = 3.4 ohm @ 5V = 1.47A or 8.5W
4 resistor = 2.6 ohm @ 5V = 1.92A or 9.6W
5 resistor = 2.1 ohm @ 5V = 2.38A or 11.9W

The first test (Picture 5) is a Viper22a SOIC-8 loaded with 3.4 ohms of resistance. I ran it for an hour (Picture 6) with the Fluke 77 set to record the minimum 5VSB voltage. I did this so that if the chip reset I would know it. It ran for the hour with no issues. The max IR temp I could get was 166 Deg F at the drain pins. If I loaded it with 2.6 ohms of resistance it would reset in a matter of just a few minutes.

The second test (Picture 7) was of the Viper27 SOIC-16 loaded with 3.4 ohms of resistance and a 100 ohm VCC resistor. It would reset in just a couple of minutes.

The third test (Picture 8) was of the Viper27 SOIC-16 loaded with 3.4 ohms of resistance and a 10 ohm VCC resistor. It would reset in just a couple of minutes.

The fourth test (Picture 9) is a Viper22a DIP-8 loaded with 3.4 ohms of resistance. I ran it for an hour (Picture 10) with the Fluke 77 set to record the minimum 5VSB voltage. It ran for the hour with no issues. The max IR temp I could get was 156 Deg F at the drain pins. If I loaded it with 2.6 ohms of resistance it also would reset in a matter of just a few minutes.

I have some Viper27 DIP I intend to test but I have to make the boards on which to mount them first.

The Viper22a chip should be able to handle more load than 8.5W so it seems something else in the system may be limiting the output when the resistance is lowered below this point.

The Viper27 SOIC-16 should be able to handle more load than the Viper22A SOIC-8 according to the datasheet but the tests do not show it. Using the 10 ohm VCC resistor did allow it to run longer than the 100 ohm but not by much.

I was actually surprised by all the results as they were not what I was expecting, especially the Viper27 tests.

[retiredcaps]

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

I ran it for an hour (Picture 10) with the Fluke 77 set to record the minimum 5VSB voltage.

1) Wow, good use of all the different types of probes and connector tips.

I think you will really like the silicone leads as they are less likely to get tangled.

2) The Fluke 12 has the Vchek (automagic detection of volts, continuity, diode, etc) feature and is sort of predecessor technology to the Fluke 113 which you hate so much.

The 12 has a min/max function that timestamps the maximum and minimum readings as per user manual snapshot attached.

3) And yes I have an used Fluke 12 and I like it. While it has the Vchek feature, I never use it. I use the manual controls instead when making measurements. The controls are slide switches and not rotary so the user interface takes a while to get used to if you are no familar with it.

[ben7]

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

According to the datasheet, the VIPer 22A soic can only do about 9-10 watts.

Do you have a snubber circuit installed? It does not look like it, from the poorly labeled circuit diagram

[LDSisHere]

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

1) Wow, good use of all the different types of probes and connector tips.

I think you will really like the silicone leads as they are less likely to get tangled.

2) The Fluke 12 has the Vchek (automagic detection of volts, continuity, diode, etc) feature and is sort of predecessor technology to the Fluke 113 which you hate so much.

The 12 has a min/max function that timestamps the maximum and minimum readings as per user manual snapshot attached.

3) And yes I have an used Fluke 12 and I like it. While it has the Vchek feature, I never use it. I use the manual controls instead when making measurements. The controls are slide switches and not rotary so the user interface takes a while to get used to if you are no familar with it.

Thanks for the information about the Fluke 12, I will keep it in mind if I decide I need another meter. I have to say I am disappointed in the spring tension of the croc clips and the way they "bite" but I do like the variety of tips I have with the TL81A kit I bought. I just need to add Fluke style connectors to my Blue ESR tester like I saw on post on the EEVblog forum.

Do you have a snubber circuit installed? It does not look like it, from the poorly labeled circuit diagram

I doubt that I have a snubber circuit since I do not know what a snubber circuit is in the first place. As far as the circuit diagram, I only copied the hand drawn schematics that Everell supplied into GEDA to make sure the connections were correct in the PCB software. The labeling is for my own purposes too suit my needs and corresponds to the pictures on how to connect this mod that Everell posted at the beginning of this thread. I am just trying to learn how these supplies work and give a little feedback on the difference in the different chips and packages. When I said the Viper22a should be able to handle more than 8.5W I was referring to the DIP package and I meant to be specific in the post but apparently I left that part out.

[ben7]

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

I doubt that I have a snubber circuit since I do not know what a snubber circuit is in the first place. As far as the circuit diagram, I only copied the hand drawn schematics that Everell supplied into GEDA to make sure the connections were correct in the PCB software. The labeling is for my own purposes too suit my needs and corresponds to the pictures on how to connect this mod that Everell posted at the beginning of this thread. I am just trying to learn how these supplies work and give a little feedback on the difference in the different chips and packages. When I said the Viper22a should be able to handle more than 8.5W I was referring to the DIP package and I meant to be specific in the post but apparently I left that part out.

Yeah yeah, thats ok, no problem!

[ben7]

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

I tried making a VIPer22A psu again but it doesn't work. With a ballast (7W lightbulb) the thing ticks, and the output pulses from about 2V to 4.5V. It is at about 1-2Hz. With no ballast (I am using a 20-watt isolation tx which limits current) the chip gets slightly hot. Tried switching all the windings on the transformer around. Both tests, the aux winding only puts out 9-12vdc (after rectification). I always have bad luck with SMPS But, I got my new oscilloscope, so I will take a look at it. (It is digital too, I could upload some snapshots of waveforms!)

[LDSisHere]

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

Ben7,

What kind of power supply are you working with? Also what kind of oscilloscope are you using, as I need to get one myself and I know very little about them?