EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: t1d on August 05, 2020, 08:11:11 am
-
I did Robert's Smorgasbord's APC Uninterruptible Power Supply Isolated Transformer Project. His YouTube Two-Part Series has all of the directions.
https://www.youtube.com/watch?v=SlTDqRNryuM&list=LLJr0cpUqRkGCBLSnwWQVPvA&index=41&t=460s (https://www.youtube.com/watch?v=SlTDqRNryuM&list=LLJr0cpUqRkGCBLSnwWQVPvA&index=41&t=460s)
Power-On Test Results that I posted in his comments.
Simple First Test Setup
Mains/Wall AC Supply = 123.1VAC.
Isolated Secondary @ No Load = 121.9VAC; a loss of 1.2VAC across the transformers.
DUT = 60W Incandescent Light Bulb. The light operated at full brightness. The isolated secondary dropped to 120.9VAC.
The Standard allowable drop for USA mains AC is 5%. For 120VAC mains, this equals 6 VAC. Therefore, 114VAC is the minimum acceptable voltage.
The unit dropped one volt, for 60 watts. This would indicate that the unit is capable of supplying (120 -114) x 60 = 360 watts. 360 watts/120VAC = 3 Amps. This is higher than my original guess-imate and I am pleased, well enough. Of course, this needs to be proven out with stress testing.
I think my model 1500 case is bigger than [the 750 case that he used.] I had plenty of room for the two transformers and a good bit of room left over, at the front and back, to add the 120V fan and the switch, fuse, sockets, wiring, etc...
I am just posting, here, to have a place for folks to see my pictures. Please forgive their poor quality. Making pictures is difficult, for me, for various reasons.
You will notice, in the picture of the front of the case, an uncovered long, vertical opening. This will be covered by a cheap multimeter, which will be permanently installed to monitor the current/voltage/etc.
-
PIX
(https://www.eevblog.com/forum/projects/diy-isolation-transformer-project-back-to-back-method/?action=dlattach;attach=1040978)
(https://www.eevblog.com/forum/projects/diy-isolation-transformer-project-back-to-back-method/?action=dlattach;attach=1040982)
-
PIX
(https://www.eevblog.com/forum/projects/diy-isolation-transformer-project-back-to-back-method/?action=dlattach;attach=1040990)
(https://www.eevblog.com/forum/projects/diy-isolation-transformer-project-back-to-back-method/?action=dlattach;attach=1040986)
-
PIX
(https://www.eevblog.com/forum/projects/diy-isolation-transformer-project-back-to-back-method/?action=dlattach;attach=1041002)
(https://www.eevblog.com/forum/projects/diy-isolation-transformer-project-back-to-back-method/?action=dlattach;attach=1040998)
-
That is a pretty good use for a pair of scrapped UPSes. I have a much bigger isolation transformer made from a pair of 1500 watt transformers which were part of old computer mass storage units.
It seems like something could be done between the low voltage windings to increase functionality but I have not come up with anything. An AC current limiter is a possibility.
-
Love the recycled pill bottle (is it a pill bottle?) for isolating the power LED stuff. The fan cut out also looks neat.
BTW I'm "Robert's Smorgasboard" (https://www.youtube.com/channel/UCGtReyiNPrY4RhyjClLifBA (https://www.youtube.com/channel/UCGtReyiNPrY4RhyjClLifBA)) ;) Maybe I should update my account on the EEVblog ...
-
This is mine ;)
[attach=1]
-
One easy way to compensate for the drop through the 2 transformers connected back to back is to connect a low voltage filament transformer in series and in phase with the output of your second transformer. This is similar to the method the self correcting autotransformers use.
If you try to make the correction between the 2 transformers you would need high current but if you add 3VAC (half the secondary of a 6.3 volt center tapped filament transformer) to the output you only need the current your second transformer can output, in your case 3A.
A 6.3V/3A c.t. transformer is fairly small and cheap and could easily be added inside your case with the other 2 large transformers and you can choose a low voltage transformer that suits your particular needs.
-
Love the recycled pill bottle (is it a pill bottle?) for isolating the power LED stuff. The fan cut out also looks neat.
BTW I'm "Robert's Smorgasboard" (https://www.youtube.com/channel/UCGtReyiNPrY4RhyjClLifBA (https://www.youtube.com/channel/UCGtReyiNPrY4RhyjClLifBA)) ;) Maybe I should update my account on the EEVblog ...
Yes, a pill bottle... I have too many empty pill bottles... Meaning, I have health challenges... I hate to trash those good bottles...
The indicator lamp is the part of the project that I think is the least clever... It is 1N4007 Diode + 22K/2W Resistor + 22K/2K Resistor + Common 5mm Red LED. The resistors were recycled, too, and their legs were not long enough to use a perf board. They are all that I had that could handle the heat, from the power drop. Even though the heat shrink extends out of the bottle, I think all the components are inside the bottle.
I coated the LED bulb with Super-Glue and Baking Soda, to create a diffused bulb. This gives a pleasant glow.
-
This is mine ;) (Attachment Link)
That is a very nice looking unit! Good job.
I see that your unit kept the front and back orientation of the original UPS... Meaning, you used the front as the front and the back as the back. So did Robert.
You will notice that I reversed the flow. This moved the power sockets and the resettable fuse to the front of the unit, for easier access.
It also changed the location of the air holes that are on the sides of the cover. The air holes are aligned with the forward transformer. This means that the fan draws air over both transformers, as the fan is in the back.
I cut out the fan hole with a side grinder and a thin metal cutting blade. I sealed the gaps with metal duct tape.
As you did, I wanted to make a professional, manufactured unit. But, making the new face plate was slowing down my progress.
The plastic sheet that isolates the circuit board from the top of the case is useful to make the face plate. I made errors with one of these sheets.
So, I used the sheet from the second unit and had it cutout perfectly. It had a slight wave in it and I thought that would make it difficult to glue it to the front of the case.
I preheated the oven, to 250*F, put it on a cookie sheet, added a book to hold it flat, turned the heat off and put it in the oven. This worked too well. The plastic shrunk, in the lateral dimension, at least 10%. It stayed long enough, though. I found that odd.
I ended up just painting it. I would still like to find something better. Live and learn.
-
One easy way to compensate for the drop through the 2 transformers connected back to back is to connect a low voltage filament transformer in series and in phase with the output of your second transformer. This is similar to the method the self correcting autotransformers use.
If you try to make the correction between the 2 transformers you would need high current but if you add 3VAC (half the secondary of a 6.3 volt center tapped filament transformer) to the output you only need the current your second transformer can output, in your case 3A.
A 6.3V/3A c.t. transformer is fairly small and cheap and could easily be added inside your case with the other 2 large transformers and you can choose a low voltage transformer that suits your particular needs.
You know so much, Authur! I always enjoy your posts.
AC is above my pay grade and it frightens me. (That fear is healthy!) That is why I built the isolation unit.
I don't exactly know how your circuit is working. Is it that the smaller transformer is simply adding current? I can surely build it, but I need to understand it. And, yes, the model 1500 case has plenty of room to add a small transformer.
-
One easy way to compensate for the drop through the 2 transformers connected back to back is to connect a low voltage filament transformer in series and in phase with the output of your second transformer. This is similar to the method the self correcting autotransformers use.
If you try to make the correction between the 2 transformers you would need high current but if you add 3VAC (half the secondary of a 6.3 volt center tapped filament transformer) to the output you only need the current your second transformer can output, in your case 3A.
A 6.3V/3A c.t. transformer is fairly small and cheap and could easily be added inside your case with the other 2 large transformers and you can choose a low voltage transformer that suits your particular needs.
Arthur, in your schematic design, there are 2 transformer cores between the mains and out adding series capacitance bridging to the mains and potentially lowering the breakdown static voltage isolation barrier, even going as far as turning the isolation to crap if that 3v transformer sucks or is defective. It would be better to add those 3v at the input to the first UPS transformers further isolating the mains and final output. Also in this configuration, the quality of that 3V transformer's isolation is now irrelevant as the 2 UPS transformers keeps you as far away from the mains as possible which is t1d's purpose in all of this.
-
One easy way to compensate for the drop through the 2 transformers connected back to back is to connect a low voltage filament transformer in series and in phase with the output of your second transformer. This is similar to the method the self correcting autotransformers use.
If you try to make the correction between the 2 transformers you would need high current but if you add 3VAC (half the secondary of a 6.3 volt center tapped filament transformer) to the output you only need the current your second transformer can output, in your case 3A.
A 6.3V/3A c.t. transformer is fairly small and cheap and could easily be added inside your case with the other 2 large transformers and you can choose a low voltage transformer that suits your particular needs.
Arthur, in your schematic design, there are 2 transformer cores between the mains and out adding series capacitance bridging to the mains and potentially lowering the breakdown static voltage isolation barrier, even going as far as turning the isolation to crap if that 3v transformer sucks or is defective. It would be better to add those 3v at the input to the first UPS transformers further isolating the mains and final output. Also in this configuration, the quality of that 3V transformer's isolation is now irrelevant as the 2 UPS transformers keeps you as far away from the mains as possible which is t1d's purpose in all of this.
Very interesting and good stuff, Brian. Please doodle me a drawing, so I am sure I understand. I am very interested to see what you and Authur come up with.
-
I re-made mine with a medical isolation transformer. It has higher noise then the shady E-core I used to use, but a isolation transformer should be safe.
-
One easy way to compensate for the drop through the 2 transformers connected back to back is to connect a low voltage filament transformer in series and in phase with the output of your second transformer. This is similar to the method the self correcting autotransformers use.
If you try to make the correction between the 2 transformers you would need high current but if you add 3VAC (half the secondary of a 6.3 volt center tapped filament transformer) to the output you only need the current your second transformer can output, in your case 3A.
A 6.3V/3A c.t. transformer is fairly small and cheap and could easily be added inside your case with the other 2 large transformers and you can choose a low voltage transformer that suits your particular needs.
Arthur, in your schematic design, there are 2 transformer cores between the mains and out adding series capacitance bridging to the mains and potentially lowering the breakdown static voltage isolation barrier, even going as far as turning the isolation to crap if that 3v transformer sucks or is defective. It would be better to add those 3v at the input to the first UPS transformers further isolating the mains and final output. Also in this configuration, the quality of that 3V transformer's isolation is now irrelevant as the 2 UPS transformers keeps you as far away from the mains as possible which is t1d's purpose in all of this.
Very interesting and good stuff, Brian. Please doodle me a drawing, so I am sure I understand. I am very interested to see what you and Authur come up with.
A doodle:
[attachimg=1]
-
One easy way to compensate for the drop through the 2 transformers connected back to back is to connect a low voltage filament transformer in series and in phase with the output of your second transformer. This is similar to the method the self correcting autotransformers use.
If you try to make the correction between the 2 transformers you would need high current but if you add 3VAC (half the secondary of a 6.3 volt center tapped filament transformer) to the output you only need the current your second transformer can output, in your case 3A.
A 6.3V/3A c.t. transformer is fairly small and cheap and could easily be added inside your case with the other 2 large transformers and you can choose a low voltage transformer that suits your particular needs.
Arthur, in your schematic design, there are 2 transformer cores between the mains and out adding series capacitance bridging to the mains and potentially lowering the breakdown static voltage isolation barrier, even going as far as turning the isolation to crap if that 3v transformer sucks or is defective. It would be better to add those 3v at the input to the first UPS transformers further isolating the mains and final output. Also in this configuration, the quality of that 3V transformer's isolation is now irrelevant as the 2 UPS transformers keeps you as far away from the mains as possible which is t1d's purpose in all of this.
Very interesting and good stuff, Brian. Please doodle me a drawing, so I am sure I understand. I am very interested to see what you and Authur come up with.
A doodle: (Attachment Link)
Thanks for the effort to provide the picture. It did help my understanding of the arrangement.
I want to make sure that I understand the theory. Is it that the additional 3V/3A transformer 1) provides headroom voltage, to prevent/limit the voltage droop and 2) additional current, for a decent power rating? Or, is it that the extra voltage and amps would just accomplish #1 (provide headroom voltage, to prevent/limit the voltage droop?) Please explain; this is beyond my learning (so far.)
I would guess that any transformer providing 3-4V and >= 3A would do. I ask if there is some leeway, because I may not have a 6VCT/3A donor transformer, in my stocks.
This is interesting = fun. Thank you for your help.
Additionally:
I found a plastic sheet that will work to fashion a face plate: https://www.lowes.com/pd/Hillman-15-in-x-19-in-Green-Plastic-Blank-Sign/3295394 (https://www.lowes.com/pd/Hillman-15-in-x-19-in-Green-Plastic-Blank-Sign/3295394)
The plastic is solid, i.e. not corrugated... Maybe 0.3-0.5mm thick. It is easily cut with a couple of passes of a razor knife. The wild, lime green color will give the case the look of the molded plastic of modern power tools, which is "cool," enough. It only cost $2.19USD and it is plenty big enough to make several revisions, should I make mistakes. Pictures to follow completion.
-
I want to make sure that I understand the theory. Is it that the additional 3V/3A transformer 1) provides headroom voltage, to prevent/limit the voltage droop and 2) additional current, for a decent power rating? Or, is it that the extra voltage and amps would just accomplish #1 (provide headroom voltage, to prevent/limit the voltage droop?) Please explain; this is beyond my learning (so far.)
Transformers always drop their output with a load on them.
That 3v addition/boost will just raise your entire system's voltage by around 3v meaning approximately 3 more volts on the output. That is unless you get the polarity backwards on the 3v transformer. Doing so means your system output voltage will be cut by 3v.
I'm not sure about your measurements, but, I would target a ~130v unloaded output. With a small 50watt load, 128v to 125v should be good as a heavy 200 watt load may load down your output close to 120v, or even a bit less. Unfortunately, you have no regulation on output voltage here.
3 amp across the full ~125v means (not counting the limits of the UPS transformers) that with that 3v @ 3a transformer in the circuit, you should not exceed ~375 watts.
-
I prefer a different method.
1kVA UPS with 24V 20A power supply installed inside instead of two lead acid 12V batteries.
You get voltage feedback, pure sine wave and at least some form of over current protection.
Think about it before you tear down another UPS unit.
That second transformer could have been used to provide the DC for another UPS.
But for many things AC is not needed.
So a small, compact flyback can be made, that supplies 200+V DC which is more than enough to power many switching power supplies.
A 50-60W supply that fits in your hand and doesn't weight a ton.
Flyback topology has pretty much a built-in overcurrent protection as most chips use current mode PWM.
Made one for supplying anode voltages for vacuum tubes and it's great.
With a half bridge topology one can go into several hundred watts.
As for back to back transformers.
It makes more sense to boost the input mains using a series transformer.
The voltage boosting transformer is on the mains side so there is still a benefit of having double isolation.
Once you insert it on the isolated mains side you are back to having single isolation.
-
I want to make sure that I understand the theory. Is it that the additional 3V/3A transformer 1) provides headroom voltage, to prevent/limit the voltage droop and 2) additional current, for a decent power rating? Or, is it that the extra voltage and amps would just accomplish #1 (provide headroom voltage, to prevent/limit the voltage droop?) Please explain; this is beyond my learning (so far.)
Transformers always drop their output with a load on them.
That 3v addition/boost will just raise your entire system's voltage by around 3v meaning approximately 3 more volts on the output. That is unless you get the polarity backwards on the 3v transformer. Doing so means your system output voltage will be cut by 3v.
I'm not sure about your measurements, but, I would target a ~130v unloaded output. With a small 50watt load, 128v to 125v should be good as a heavy 200 watt load may load down your output close to 120v, or even a bit less. Unfortunately, you have no regulation on output voltage here.
3 amp across the full ~125v means (not counting the limits of the UPS transformers) that with that 3v @ 3a transformer in the circuit, you should not exceed ~375 watts.
Very interesting, Brian. Thank you for teaching me about this. If I find that I need more power support, I will know what to do.
-
I prefer a different method.
1kVA UPS with 24V 20A power supply installed inside instead of two lead acid 12V batteries.
You get voltage feedback, pure sine wave and at least some form of over current protection.
Let's make sure I understand... You are suggesting to leave the UPS in its original form and substitute transformers for the batteries. I take it that you would mean that this source would be rectified and regulated to 24VDC. Then, when isolation is needed, power the DUT off of the UPS... Correct? Meaning the inverted signal from the UPS would be isolated from the mains? If so, that is another clever approach, if you only have one donor UPS and a donor 24V/20A transformer.
As for the output of such a configuration being a pure sign wave... I would think that feature would be lost, because it would be achieved from the batteries being the source. Once the 24V/20A transformer was added, isn't that lost? I don't know; as said, AC is outside my knowledge base...
But, an interesting idea, Pawelr. Thanks for joining in.
-
I prefer a different method.
1kVA UPS with 24V 20A power supply installed inside instead of two lead acid 12V batteries.
You get voltage feedback, pure sine wave and at least some form of over current protection.
Let's make sure I understand... You are suggesting to leave the UPS in its original form and substitute transformers for the batteries. I take it that you would mean that this source would be rectified and regulated to 24VDC. Then, when isolation is needed, power the DUT off of the UPS... Correct? Meaning the inverted signal from the UPS would be isolated from the mains? If so, that is another clever approach, if you only have one donor UPS and a donor 24V/20A transformer.
As for the output of such a configuration being a pure sign wave... I would think that feature would be lost, because it would be achieved from the batteries being the source. Once the 24V/20A transformer was added, isn't that lost? I don't know; as said, AC is outside my knowledge base...
But, an interesting idea, Pawelr. Thanks for joining in.
Yeah, power the DUT from the UPS, just disconnect the internal ground going to the output socket to make it fully isolated.
In my case I installed a 24V 20A switching power supply inside the UPS.
I got it for free, as well as the UPS. They were both headed for the scrapyard.
It supplies fairly well regulated voltage for the UPS to use.
Even under load the output sine wave was OK.
But another transformer may be used to get the DC voltage too.
Anywhere from 22V to 30V should be fine.
Will it handle the rectified 24Vac (30+V) ? To answer that I would have to look up every chip in the UPS to check if they can withstand such voltage or if it has any overvoltage protection implemented.
The UPS will have some kind of feedback so the ripple should be canceled.
How well would it work is another question.
At least in my case it does react fairly well to load or input voltage changes.
Non linear loads (rectifier + capacitor) do distort the sine wave a little (nothing serious), which is expected.
-
This is the plastic I used to make the replacement face plate. I think it has the look of the cool, modern molded plastic tools... Like Ryobi, or Makita... I may add some labeling...
https://www.lowes.com/pd/Hillman-15-in-x-19-in-Green-Plastic-Blank-Sign/3295394 (https://www.lowes.com/pd/Hillman-15-in-x-19-in-Green-Plastic-Blank-Sign/3295394)
It is thin, solid material. Not corrugated. Easily cut with a razor knife. Just $2.18.
(https://www.eevblog.com/forum/projects/diy-isolation-transformer-project-back-to-back-method/?action=dlattach;attach=1047070)
-
Interesting thread. I went shopping at my favourite store today, "La Strada" or the street and found a pair of identical UPS's. What are the chances you say? Not sure but clearly not zero.
I even found a lead acid 900A car booster jump starter. Pressed the battery check button and it was fully charged. Just need to find an adapter to keep it charged. I wonder if the battery will fit in the UPS?
-
The idea of using two transformers in series (back to back) is, in my opinion, not good because it has drawbacks.
Indeed, the resulting short-circuit impedance is doubled, which potentially can perturb the operation of the DUT.
Let's remember the equivalent diagram of the transformer:
Two virtual inductors exist, lf1 at the primary and lf2 at the secondary as well as two resistors, r1 due to the ohmic resistance of the primary and r2 due to the ohmic resistance of the secondary.
This set of resistances and inductances is measured by a transformer short circuit test.
A reduced voltage is applied to the transformer primary with a variac, secondary in short circuit.
When the secondary current is equal to the nominal secondary current, we denote the primary voltage which is x% of the nominal primary voltage.
The short circuit impedance is then x%.
This is more or less 4-5% for a toroidal transformer, 7-8% for an EI transformer and about 15% for a MOT transformer without magnetic shunt and 25% with it.
The short circuit impedance of two transformers back to back is the sum of the short circuit impedances of the two transformers, which is equivalent to putting in series with the DUT an inductive impedance which can alter the operation of the DUT.
For example, the shape of the current drawn by an SMPS (without PFC) will be different from that without this isolation transformer device.
This is why it is advantageous to use an isolation transformer with the lowest possible short circuit impedance, which makes it necessary to choose a toroidal transformer.
Personally, I use a 1KVA toroidal transformer for medical use, with reinforced 5KV primary / secondary insulation.
-
The idea of using two transformers in series (back to back) is, in my opinion, not good because it has drawbacks.
Indeed, the resulting short-circuit impedance is doubled, which potentially can perturb the operation of the DUT.
I usually consider that an advantage because it limits the current and better represents the worst case conditions however larger transformers than strictly necessary can be used if the increased impedance is a problem.
-
The peak and the rms values of inrush current of an SMPS are very different with or without this additional inductor.
A SMPS power supply could therefore burn its input fuse without an isolation transformer, but not with it.
The operation is therefore very different, especially with regard to the inrush current and the waveform of the current consumed.
It is therefore not an advantage in terms of measurements and tests.
If you want to limit the current, use an external device like a 100W incandescent bulb in series with the DUT
-
I just wanted to give an update on using the isolation transformer, over time.
- Firstly, it does what it was intended to do - keep me safe(r.)
- Secondly, I think it can handle about 3A/360W. That has been sufficient for my needs.
- Thirdly, the case is large and bulky. That was convenient for building it = plenty of room to work in. But, given its big footprint on my bench, I would look for some way to make it smaller. Two ways that might work... A) I reused the existing socket strip. Reduce that to a single socket. B) Stack the transformers vertically.
Hope that helps.
-
I know you want to use the transformers from the UPSs, but another solution if you can get hold of two 240V from/to 120v conversion transformers, is connect the primaries in parallel, and the secondaries in series. That way each transformer carries only half the load.
(Or get four 30v transformers and do that!!??)
A couple of threads I started may be of interest:-
https://www.eevblog.com/forum/projects/back-to-back-isolation-transformers-now-in-a-box-plus-other-contraptions/msg2783828/#msg2783828 (https://www.eevblog.com/forum/projects/back-to-back-isolation-transformers-now-in-a-box-plus-other-contraptions/msg2783828/#msg2783828)
The above also refers to this:-
https://www.eevblog.com/forum/projects/testing-back-to-back-power-transformers-for-use-as-an-isolation-transformer/msg2398320/#msg2398320 (https://www.eevblog.com/forum/projects/testing-back-to-back-power-transformers-for-use-as-an-isolation-transformer/msg2398320/#msg2398320)
-
I know you want to use the transformers from the UPSs, but another solution if you can get hold of two 240V from/to 120v conversion transformers, is connect the primaries in parallel, and the secondaries in series. That way each transformer carries only half the load.
(Or get four 30v transformers and do that!!??)
A couple of threads I started may be of interest:-
https://www.eevblog.com/forum/projects/back-to-back-isolation-transformers-now-in-a-box-plus-other-contraptions/msg2783828/#msg2783828 (https://www.eevblog.com/forum/projects/back-to-back-isolation-transformers-now-in-a-box-plus-other-contraptions/msg2783828/#msg2783828)
The above also refers to this:-
https://www.eevblog.com/forum/projects/testing-back-to-back-power-transformers-for-use-as-an-isolation-transformer/msg2398320/#msg2398320 (https://www.eevblog.com/forum/projects/testing-back-to-back-power-transformers-for-use-as-an-isolation-transformer/msg2398320/#msg2398320)
Thank you for your participation and contribution. Excellent! I haven't read the threads, yet, but I like the idea of using multiple transformers, if that is what a person has at hand. They could add switches and vary the DUT voltage, by turning on/off additional transformers.