Question on Power Supply Diodes

[Electro Fan]

I'm guessing there might be a 20V version that might address floobydust's concern too.

Consider the 16.5 V, I linked to above.

Unloaded, it is probably 20 V or more, because transformers tend to increase their output voltage, when they have little/no output current (Transformer's Load Regulation).

(Rough rule of thumb, before voltage losses, such as diodes), gives approaching 30 V, but the circuit also has 5.1 V (see original schematic) on the other rail, giving around 35 V, minus diode drops.

But the op-amps (depending on what you have exactly), may only be specified to 36 V.

So, maybe 16.5 V (AC), is around the practical limit, for this exaggerated spec power supply kit.

(Including diode drops, could increase that to 18 V AC, but on the other hand, mains voltage, can fluctuate, so best to have a safety margin).

tl;dr
Even 20 V AC, could be too high, when considering transformer load regulation, and possible mains voltage fluctuations.

On the other hand, you could be very lucky, and have the 44 V rated op-amps, people talk about, but it is not clear, what you have.

EDIT:
I can believe/agree, that if you carefully checked the components you have in the kit (especially what op-amps they really are), accurately calculated all the voltages (including diode drops etc), that some AC figure of around 20 V (AC), would be ok (with the possible exception of mains voltage fluctuations).
But, I feel more comfortable, with a bit of safety margin, hence the lower transformer V AC recommendations, plus your supplier doesn't seem to do ones of 20 V AC, that I could see (pre-cased ones).

Per your earlier analysis, here is some Q&A from the comments section of the VoltLog video:

2 months ago
hi can I use a 24v 5a transformer for the above circuit, since i have it with me. thanks

1 week ago
yes you can, but you should not. .  The opamps have a maximum voltage of 30v.  After filter, a voltage of almost 34 volts apear. Under load, I'm sure it is not a issue, but I let mine powered on over night with no load and burn the opamps. This power supply shoud be ratated maximum to use with a 20v transformer. 12 0 12 or 24 0 is too much for this supply.

Looks like your 16.5V recommendation should be a good starting point.  Thx again for the guidance.

[Electro Fan]

One really good breadcrumb leads to the next.

The VoltLog power supply enclosure:

[Electro Fan]

Still waiting on a couple transformers but the rest of the kit is pretty much ready to go.  Going to start with a 16.5V 2400mA transformer as recommended by MK14.

In the meantime I noticed that the fan on the heat sink is powered by a 24V header.  I tested the fan and found that it would turn on around 6V and about 10mA.  At 12V it draws 41 mA, at 18V it draws 58 mA, and at 24V it draws 77 mA.  I'm hoping their is some current left in this design to power something more than the fan. 

[ledtester]

A TO-220 LM78xx regulator is good for at least 1 amp provided it is adequately kept cool. You can always substitute different linear regulator to drive the fan and other components - like this guy did:

https://www.qsl.net/z33t/dc_0-30v_0-3A_eng.html

[MK14]

Still waiting on a couple transformers but the rest of the kit is pretty much ready to go.  Going to start with a 16.5V 2400mA transformer as recommended by MK14.

In the meantime I noticed that the fan on the heat sink is powered by a 24V header.  I tested the fan and found that it would turn on around 6V and about 10mA.  At 12V it draws 41 mA, at 18V it draws 58 mA, and at 24V it draws 77 mA.  I'm hoping their is some current left in this design to power something more than the fan. 

That is a good idea (if I say so myself).
As others have suggested, it is possible to redesign it (partly), and change various bits, so that it can cope with high voltages. But, it is not so easy to perform all the modifications and get hold of all the components, etc.

You MIGHT get away with higher voltages. But at some point, it would either break straight away, and/or be unreliable and/or have significantly shorter component life.

You will probably get something like 19 V AC from a transformer like that (16.5 V AC), when it is very lightly loaded. Because transformers, have a load regulation voltage parameter, which means it goes up by a percentage (e.g. 10% or 15% etc, see datasheet and/or spec sheets, if available), when lightly loaded (less than rated max current, is used).
It's proportional to the current used, so at near 0 amps, will be at its highest.

[Electro Fan]

Still waiting on a couple transformers but the rest of the kit is pretty much ready to go.  Going to start with a 16.5V 2400mA transformer as recommended by MK14.

In the meantime I noticed that the fan on the heat sink is powered by a 24V header.  I tested the fan and found that it would turn on around 6V and about 10mA.  At 12V it draws 41 mA, at 18V it draws 58 mA, and at 24V it draws 77 mA.  I'm hoping their is some current left in this design to power something more than the fan. 

That is a good idea (if I say so myself).

LOL - that's funny

As others have suggested, it is possible to redesign it (partly), and change various bits, so that it can cope with high voltages. But, it is not so easy to perform all the modifications and get hold of all the components, etc.

You MIGHT get away with higher voltages. But at some point, it would either break straight away, and/or be unreliable and/or have significantly shorter component life.

You will probably get something like 19 V AC from a transformer like that (16.5 V AC), when it is very lightly loaded. Because transformers, have a load regulation voltage parameter, which means it goes up by a percentage (e.g. 10% or 15% etc, see datasheet and/or spec sheets, if available), when lightly loaded (less than rated max current, is used).
It's proportional to the current used, so at near 0 amps, will be at its highest.

If you can explain more about this or provide a link to something tutorialish I'd like to better understand this.  Thanks

[Electro Fan]

A TO-220 LM78xx regulator is good for at least 1 amp provided it is adequately kept cool. You can always substitute different linear regulator to drive the fan and other components - like this guy did:

https://www.qsl.net/z33t/dc_0-30v_0-3A_eng.html

Thanks!  That's another very good link about this power supply project.  The schematic helped me understand a little more than the previous schematic I had. 

I'm wondering if the op amps are easy low hanging fruit that could be the beginning of an upgrade path?  Ultimately it would be cool to get the whole thing up a notch or two on voltage and current but especially current.

[MK14]

If you can explain more about this or provide a link to something tutorialish I'd like to better understand this.  Thanks

Assuming you mean about the transformers, output voltage changes (regulation):

https://www.electronics-tutorials.ws/transformer/voltage-regulation.html

https://www.allaboutcircuits.com/textbook/alternating-current/chpt-9/voltage-regulation/#:~:text=Voltage%20regulation%20is%20the%20measure,the%20regulation%20it%20will%20provide.

[Electro Fan]

If you can explain more about this or provide a link to something tutorialish I'd like to better understand this.  Thanks

Assuming you mean about the transformers, output voltage changes (regulation):

https://www.electronics-tutorials.ws/transformer/voltage-regulation.html

https://www.allaboutcircuits.com/textbook/alternating-current/chpt-9/voltage-regulation/#:~:text=Voltage%20regulation%20is%20the%20measure,the%20regulation%20it%20will%20provide.

You are an excellent curriculum finding teacher !  Those articles hit the spot.

Just to confirm, the second article indicates that 3% voltage regulation might be relatively "good" and that ~6.9% might be relatively "poor".  Are these reasonable bookends on some scale or for some basic use case?  Thanks again

Edit:
Looks like this (maybe similar to the 7824 used in the PS project) might come in versions that range from 1.5% to 4%, so presumably this is in the good (enough) range for various applications.... ?

https://www.mouser.com/datasheet/2/308/MC7800_D-1773680.pdf

[Electro Fan]

If you can explain more about this or provide a link to something tutorialish I'd like to better understand this.  Thanks

Assuming you mean about the transformers, output voltage changes (regulation):

https://www.electronics-tutorials.ws/transformer/voltage-regulation.html

https://www.allaboutcircuits.com/textbook/alternating-current/chpt-9/voltage-regulation/#:~:text=Voltage%20regulation%20is%20the%20measure,the%20regulation%20it%20will%20provide.

You are an excellent curriculum finding teacher !  Those articles hit the spot.

Just to confirm, the second article indicates that 3% voltage regulation might be relatively "good" and that ~6.9% might be relatively "poor".  Are these reasonable bookends on some scale or for some basic use case?  Thanks again

Edit:
Looks like this (maybe similar to the 7824 used in the PS project) might come in versions that range from 1.5% to 4%, so presumably this is in the good (enough) range for various applications.... ?

https://www.mouser.com/datasheet/2/308/MC7800_D-1773680.pdf

And this one (in version A) is 2%
https://www.mouser.com/datasheet/2/389/l78-1849632.pdf

[MK14]

Unfortunately, there seems to be some confusion. I need to take some/all of the blame, so don't feel bad.
The term "Voltage Regulation", can be confusing, as it seems to be used, for a number of things.

I was referring to Transformer "Voltage Regulation", which in very simplistic terms just means the voltage varies by say 10% (approx), as the output (load) current changes.
My concern was that it could make the transformer output voltage (AC) too high, for that PSU. Hence talking you into using a lower transformer rating (which you have), which solves that problem.

What you have just linked to, is another form of "Voltage Regulation", which is integrated circuit voltage regulator's "Voltage Regulation". Which also can vary with output (load) current changes.
Maybe a few percent or so, on a cheap/simple regulator (varies widely).

The idea is that the final voltage regulator (which is most of the PCB and circuit, you just bought), but can be just a simple 3 lead, integrated circuit regulator, just like you linked to. Will attempt to filter out (ignore), the transformers voltage variations, voltage ripple and things, to give a nice clean, well regulated output voltage.
BUT, it is not going to be perfect, and will also vary by a percentage (hopefully rather small), as the load (current) increases, temperature changes, and other things.

[MK14]

tl;dr (Summary):

Transformer "Voltage Regulation", just the raw transformer, no extra circuitry or ICs. How its AC voltage drops/changes with transformer output current into the load (of the transformer).

(What you seem to be thinking of). "Voltage Regulation" of the voltage regulator(s), which is something else. It is how stable the voltage regulators output voltage is, with changes in output load current.

[Electro Fan]

tl;dr (Summary):

Transformer "Voltage Regulation", just the raw transformer, no extra circuitry or ICs. How its AC voltage drops/changes with transformer output current into the load (of the transformer).

(What you seem to be thinking of). "Voltage Regulation" of the voltage regulator(s), which is something else. It is how stable the voltage regulators output voltage is, with changes in output load current.

Thanks for the clarifying posts.  No worries.

Your links were helpful and I think I get some of the basics:  A transformer's output by itself can vary based on load or no load - perhaps by 10% or so, hence your margin-adding recommendation to reduce the spec'd voltage to around 16V.  Additionally a voltage regulator can improve the performance of the overall circuit by reducing the variation to 2% or less.  It's all good - and between building the PS and doing the Q&A here it's coming into better focus. 

Pretty soon I'm going to plug in the AC and see what happens on the DC output

Thx again

[Electro Fan]

New question:  I now have a 24V 2A transformer.  When plugged into my ~120V AC (which at the moment measures 123.4V with no load - normally I have seen about 121V) the transformer output measures 28.2V with no load.  What do you think, safe to give it a whirl with the PS or not enough margin for error?  Thx

Update:  I have a Dale 50 Watt 50 Ohm 1% resistor (that measures ~50.2 Ohms).  I'm thinking if I power up with the PS loaded by the resistor, maybe it's going to be ok if I put the PS together properly...?

Update 2:  Or are you saying that voltage seen by the op amps is for some reason independent of the voltage reduction that occurs when PS is loaded?

[MK14]

New question:  I now have a 24V 2A transformer.  When plugged into my ~120V AC (which at the moment measures 123.4V with no load - normally I have seen about 121V) the transformer output measures 28.2V with no load.  What do you think, safe to give it a whirl with the PS or not enough margin for error?  Thx

Update:  I have a Dale 50 Watt 50 Ohm 1% resistor (that measures ~50.2 Ohms).  I'm thinking if I power up with the PS loaded by the resistor, maybe it's going to be ok if I put the PS together properly...?

Update 2:  Or are you saying that voltage seen by the op amps is for some reason independent of the voltage reduction that occurs when PS is loaded?

The capacitor(s), immediately after the bridge rectifier, will attempt to reach almost 1.5 (root 2) times the peak AC voltage (28.2 V, peak as in unloaded), so they would reach 28.2 x root 2 - 1.2 volts worth of diodes =
= 38.6 V plus the other 5 V rail = almost 44 volts (peak, when unloaded).

44 V is the absolute maximum voltage on even the much high voltage op-amps, which YOU DO NOT HAVE, you have the 36 V ones, which are only really specified (working), at 30 V.
Too high a voltage, could instantly fry all your op-amps.

Solutions:
1) Just connect it up. Not recommended.

2) Wait for the new, lower voltage transformer. Probably the best thing to do.

3) Drop the voltage, using other ways. Such as diodes in series. Or putting a load on the dropping resistor (50 ohms).

N.B. Once you over-voltage those op-amps, they may instantly break and never work again. Causing major headaches, when you try and fix or get this item to work.

tl;dr
This is getting so complicated (because of various misunderstandings), and trying to rush things.

Best to start at low voltage and low current. Get it to work, then go from there.

[Electro Fan]

Hi MK14,

Thank you very much for your guidance.

Turns out that I went with your Solution 2.  I used an enclosed transformer spec'd for 16V and 2.5A.

I am very happy to report that the power supply started up on the first try and ran through multiple tests - probably for nearly an hour.

Attached below are the measurements.

A few key results:

With a 50 Ohm resistor the maximum current output was 381 mA with a max DC voltage of 19.1

With a 1 Ohm resistor the PS achieved a max current of 2.7 Amps and a max voltage of 2.7; however I noticed that this put a load of 3.9 Amps on the secondary side of the transformer so I didn't run those settings very long.

At approximately 1V and 1A DC (0.98V and 0.98A) the power supply reached 1.75A on the secondary at 16.97V

I scientifically tested temperature with my fingers and found everything ran cool on all tests with the exception of the 5W 47 Ohm resistor right in front of D1047.  That resistor might be my first thing to upsize - maybe to 50 watts.  It was too hot to touch for more than a few seconds, but to be square I think I noticed it when I was pushing the PS at it's max output pf 2.7A and 2.7V (which is when the secondary was beyond it's spec at 3.9A).

Throughout the test the small fan on the big heat sink at D1047 operated like a champ, adjusting speed based on the V and A potentiometer settings.  Most of the time it sounded like a small jet engine (this is not something anyone who complains about commercial test equipment fans would find acceptable - it's louder than all the fans on my bench combined - but it seemed to do it's job admirably).

I managed to get an on-off switch inline between the transformer secondary and the PS PCB.  I'd recommend this to anyone testing as an initial step.  Longer term I'll get a power switch on the primary side of the transformer.

Your suggestion to use the enclosed transformer and clearly your overall sizing recommendations were very sound advice.  Thanks!

All in all, for about $11 including the PCB and components with the big heat sink and fan, plus another $13 for the transformer it's a highly recommended project as a learning experience.  I wouldn't recommend this as an alternative to an off the shelf PS for home lab work but for entry level learning about power supplies I'd rate it an A if not an A plus.

Not sure what's next.  Probably some more tests,  Maybe building in some metering (vs 4 DMMs).  And maybe/possibly an enclosure.

[MK14]

Congratulations!   

I'm glad and pleased, that worked out rather nicely for you.

Thanks for coming back, and giving such a nice, detailed report on your successes.
Many posters, don't do that. But, such feedback is useful for all the contributors to this thread, and a useful source of information, for anyone who reads it in the future.

[Electro Fan]

Ok, back at it.

In case anyone wants to try the same transformer, it's a MG Electronics Model #MGT1640.
https://www.jameco.com/shop/ProductDisplay?catalogId=10001&langId=-1&storeId=10001&productId=2230635
It says 16.5V, 2400mA in the link but on the box it says 16V, 40 VA.  The one I received has a nice green power light.

Questions...

It might make sense to replace the 5 watt 47 ohm resistor with one of these:
https://www.amazon.com/gp/product/B07JN3BFNN/ref=ox_sc_act_title_7?smid=A1OAXIQ154HL11&psc=1

but, assuming I could make it fit/attach, I'm wondering if this would buy a bunch more head room:
https://www.amazon.com/gp/product/B07JN3GF4G/ref=ox_sc_act_title_9?smid=ATNHGLND9W21T&psc=1

Also, I saw there was another thread going on about op amps but just checking to see if anyone has a favorite op amp that would be recommended to gain some more headroom over the TL081CPs?

[Electro Fan]

Not a lot of answers to those last questions, but no worries.... here's a new question (or two):

If the power supply (under discussion in this thread) is set for 1.5 Volts and 1.5 Amps into a 1 Ohm load, a) roughly what temperature might be expected at D1047 and b) according to this spec sheet what would you expect is the maximum safe operating temperature for D1047?

https://www.st.com/resource/en/datasheet/2sd1047.pdf

Thanks

[bdunham7]

The datasheet lists T_{j max} at 150C.  That's the limit and typically you try to stay well under it.

You don't actually have enough info to do the math here--you don't know what the rectified/filtered pre-regulator voltage is, but given a 16V 2.5A transformer and a 1.5A load, I'll guess it is 18 volts.  And you don't know the specification of the heat sink/fan combo, but let's guess and say 3K/W with the fan at 18 volts (meaning the heat sink will be 3 degrees C/K higher than ambient per watt being dissipated).

So, now you have a voltage drop across the transistor of 16.5 volts and a current of 1.5A, so 24.75 watts, let's round that to 25 watts.  The transistor datasheet lists the thermal resistance as 1.25K/W, so assuming the transistor is firmly attached directly to the heatsink with good thermal compound, the overall package is about 4.25K/W.  25 watts x 4.25K/W = 106.25 degrees C/K of temperature differential.  Assume an ambient temp of 25C and you have T_j of 131.25C.  That's pretty toasty, you certainly don't have much margin left if either of my two guesses is off the wrong way.

[Electro Fan]

The datasheet lists T_{j max} at 150C.  That's the limit and typically you try to stay well under it.

You don't actually have enough info to do the math here--you don't know what the rectified/filtered pre-regulator voltage is, but given a 16V 2.5A transformer and a 1.5A load, I'll guess it is 18 volts.  And you don't know the specification of the heat sink/fan combo, but let's guess and say 3K/W with the fan at 18 volts (meaning the heat sink will be 3 degrees C/K higher than ambient per watt being dissipated).

So, now you have a voltage drop across the transistor of 16.5 volts and a current of 1.5A, so 24.75 watts, let's round that to 25 watts.  The transistor datasheet lists the thermal resistance as 1.25K/W, so assuming the transistor is firmly attached directly to the heatsink with good thermal compound, the overall package is about 4.25K/W.  25 watts x 4.25K/W = 106.25 degrees C/K of temperature differential.  Assume an ambient temp of 25C and you have T_j of 131.25C.  That's pretty toasty, you certainly don't have much margin left if either of my two guesses is off the wrong way.

Hi bdunham7,

Thanks for the excellent reply.  Not only did you provide the answers but also the methods for arriving at the answers - that was great!  (And fast too.) 

I attached the transistor to the heat sink with with I think is pretty good thermal compound.  So far when I keep the AC current below 2.5 Amps (which is easy to do when I keep the DC below 1.5 V and 1.5 A into the the 1 Ohm load) the max temp I've seen on D1047 is about 48 C (in a room with an ambient temp of 23 C) ..... so maybe I can quit worrying about the temp?

I consistently see about 16.5 V on the AC side of the transformer when I've been trying to keep it under 2.5 Amps on the AC side.  The AC voltage doesn't change much based on what happens on the DC side but it's not hard to push the transformer's AC to 3 Amps or more. At 2 Volts and 2 Amps DC the transformer AC Voltage drops to about 15.7 V and the transformer current goes to about 3.05 Amps, which runs D1047 to about 60 C.  At this point the big cap gets to about 40 C but the op amps and the big resistor barely get beyond the ambient temp.   

So far it seems that the main thing I need to keep an eye on as I dial up more DC is the current on the AC side of the transformer.  If that was spec'd for more headroom I think everything else might be able to perform up a notch.  To be square, I don't really understand exactly what's causing what but the more I measure the more I see some of the relationships, and your explanations on the temp helped a bunch as along with the AC current the temp seems to be a very  key thing to keep an eye on.

Thanks again for walking me through the processes!

[bdunham7]

Do keep in mind that you can only measure the case temp (not the black plastic, but the metal tab), not T_j.  So you have to calculate the wattage, which is the voltage dropped across the transistor x the output current, and multiply that by 1.25 and add that number to the measured case temp to get an approximation of T_j.  So if at 2 amps, 2 volts with a pre-regulator 16V, you have 14V x 2A = 28 watts x 1.25 = 35K difference between T_j and T_case.  So if you are measuring 60C at the case, the chip is at 95C.  Your heat sink must be much better than my guess, b/t/w.

And just an additional note, measuring the AC voltage of a transformer that is feeding a rectified/filtered DC supply is problematic because there will be severe waveform distortion.  Yes, a true RMS meter will correctly read the RMS value, but your circuit isn't responsive to the RMS value, it is only responsive to that portion of the cycle that exceeds the voltage on the filter cap by 2V_f of the rectifiers.  That all gets complicated quickly so I'll simply sat that it sounds like you have a nice 1.5 amp power supply.

[MK14]

At 2 Volts and 2 Amps DC the transformer AC Voltage drops to about 15.7 V and the transformer current goes to about 3.05 Amps, which runs D1047 to about 60 C.

The 60 deg C, could easily be too much temperature for the power transistor, let's do a quick rough calculation.

If there is (16.5 x 1.414 - 2V (diodes)) - 2V output = 19.3 volts and 2 amps across the power transistor, then it would be dissipating around 38.6 watts.

If the transistors junction to case is 1.25 deg C per watt + something like 1 deg C per watt of heatsink to case, thermal resistance.

Then (rough back of envelope calculation), = 2.25 Deg C per watt x 38.6 watts = 60 deg C plus 87 deg C = around 147 deg C = on the high side, as it is really/ideally spec'd for 125 deg C (The 150 deg C mentioned earlier, is the absolute maximum, but it is basically spec'd for a recommended max of 125 deg C). But the datasheet does seem to allow use up to 150 deg C internal JUNCTION temperature.

My figures a bit too high, for two reasons. One is that the ripple voltage will tend to reduce the dissipation (I've ignored ripple), and you may have succeeded in getting better than 1 deg C per watt, between the transistors case and the heatsink.

[bdunham7]

My figures a bit too high, for two reasons. One is that the ripple voltage will tend to reduce the dissipation (I've ignored ripple), and you may have succeeded in getting better than 1 deg C per watt, between the transistors case and the heatsink.

I was going to say...

The filter cap voltage might be 21-22 volts no load, but there's no way it is that high at full load and you can't use his AC voltage measurement x 1.414 to determine it--it will be  much lower.  Also, I had understood that he was measuring T_case, so that the case-to-heat sink junction wouldn't be included in the T_j calculation.  And yes, 1K/W would be a crummy installation, should be a third of that or less.

[MK14]

I was going to say...

The filter cap voltage might be 21-22 volts no load, but there's no way it is that high at full load and you can't use his AC voltage measurement x 1.414 to determine it--it will be  much lower.  Also, I had understood that he was measuring T_case, so that the case-to-heat sink junction wouldn't be included in the T_j calculation.  And yes, 1K/W would be a crummy installation, should be a third of that or less.

I agree, the voltages will usually be lower, than I stated, because it will drop a fair bit at full load. In fact, I think he is actually overloading the transformer, at that point. So, the voltages will tend to be even lower, still.

On re-reading it (I incorrectly thought he had specifically said it was the heatsink temperature he was measuring).