last nights attempt at M51 with 1100D

[DrRobin]

Hi,

Thanks for the tip about the number of darks/flats/bias, looks like I am going to have to do some more in future. I suppose I could run some of these at the start of the session whilst I am waiting for it to get dark?

Thanks for the idea of a flat light box. I have seen EL panels in A3 for around £70, but then import duty will take that up to around £85, possibly more. They are available in the UK, but price is £100 + VAT + delivery. I also found some LED panels which are meant for posters in the UK at around £60 for A3 ( http://www.ledsignsandlighting.co.uk/light-sheet-aluminium-frame-p-160.html ), but not sure how uniform they might be, not sure if anyone has tried one?

I am tempted to buy one and try it. If it's no good then I would put it on flea bay for someone to put a poster in it. At that price it will be a lot less bother buying and cutting perspex.

[obscura]

The flats are quick to do and, if the scope is left as is, could be done the following day as they are not temperature critical. The darks/bias must be done at same temp as subs as we know. I prefer to do the darks (and flats and bias) direct after subs after stopping track but you're right, a lotta time waiting. Give it a go, nothing to lose but an earlier night to bed to gain.

To be honest, I do little during these long spring/summer nights. I just don't have the wherewithal to hang around waiting for dark only to find that haze or thin cloud waiting for me. But I will be out there to catch the odd one here and there as they drift into a nice piece of sky.

Sorry, I don't know of these panels. I have a table saw and cutting the perspex took seconds - two equal rectangles and they're done. Think my flat box was less than a tenner which is good for my 80 and 120mm fracs. But I guess that'salot different to 8 and 11".

[mindburner]

Hi,

I would have a look at an LM338T (or similar). It is capable of 5A (7A peak), but depends on your voltage differential. It's fully contained and is thermally and over current protected.

If you need a total load of more than 5A, just add a couple of devices in parallel, but don't common the outputs, just have separate feeds of 12V, 5A each. Best of all they are cheap (sub £2).

Seperate feeds would be good as far as noise and stability is concerned, run your mount off one feed, the Peltier cooler off another and your imaging cameras/guiders off a third.

You can keep your toroid, rectifier and smoothing circuits and there isn't much to connect after that. No power resistors to warm up and less thermal issues, just mount the regulators on a good heatsink and away you go.

BTW, a metal block isn't always a good heatsink. It can soak up a lot of heat, but if it can't dissipate it to the air, it just gets warmer and warmer, until you go over maximum temperature. A heatsink with fins is the best, these are probably easily available as add ons for fast processors and will have a fan.

My first stack of M51 (25 out of 60 subs) was a little lacking in detail, DSS chose a lot of frames from earlier in the evening where there was a lot of light in the sky. I am currently re-stacking, forcing it to use the later frames.

hi

I managed to replace the faulty transistors and got the PSU running. At least the voltage selection is working but the current limiting pot has no effect at limiting the output. I like the idea of using regulators but I need one single meaty output as I want this for a bench PSU rather than for the mount / scope. I have a 5A switch mode for this already.

Would you have any idea why the current limiting is ineffective. I am starting to think the original circuit may have issues.

I really should start afresh with the electonics but I have gone so far with this and have the transistors and heat sink etc already.

thanks again

[DrRobin]

Hi,

Looking at the circuit and datasheet the current limit should work. With the 500ohm pot at the bottom the current limit with the resistor values would be 3A. This would increase as the pot is turned up to no current limit at all.

Current limit will start to work when the voltage on pin 2 is around 0.7 volts greater than pin 3. The difference in circuit is two 0.47 ohm resistors in parallel, which gives 0.23 ohms. You could try measuring this with a meter or scope to see if it ever reaches 0.7 volts and at what current.

From ohms law v = ir, the current in a 0.23 ohm resistor to give 0.7 volts is 3A.

If you want a smaller current limit, increase the value of the two 0.47 ohm resistors, although this might reduce line regulation a little bit.

If it is not limiting at 3A minimum then check your circuit to make sure everything is connected correctly as the circuit diagram looks correct.

I have had another look at the transistors in parallel. The 0.1 ohm resistor in the emitter should provide stability so that each transistor doesn't supply too much current. However, at 0.1 ohms it will probably only make a difference above around 3 amps, well within the spec of a 2n3055. However, if you want them to share the load better, increase this resistor slighty, perhaps 0.3 ohm, but it all depends on total output current.

The other way of stabilising the circuit is to add base resistors between the BD139 and the 2n3055. A value of around 10 ohms will limit base current and make sure you don't over drive one of the 2n3055s. How much good this will do is debatable.

Let me know how you get on.

[DrRobin]

Hi,

I had a couple of other thoughts, if you measure the voltage on the emitter of the 2n3055 then it is possible to work out if each transistor is sharing the load. If the load resistors are the same value, then if the transistors are taking the same load then the emitter voltages should be the same.

I think you mentioned that the output voltage is 12v, but can you measure the input voltage and the maximum load current you are going to pull (15A?). I just want to work out the power dissipation.

If you let me know your heat sink size, think it is a block of aluminium then I can work out the thermal efficiency, just to check it isn't going to overheat.

Finally, I had one further thought about the transistors in parallel, the emitter resistors should provide the balance but a better way would be to control the first transistor using the LMxxx device and then use Current Mirrors to slave the other 2n3055s from the first.

A current mirror balances the current flowing in two circuits regardless of the differences in the transistor characteristics, if you have problems with the existing arrangement I can sort a circuit out for you, they are not complex.

[mindburner]

hi

thanks for all your input into this. I managed to get the current limiting working. Basically I had bypassed the 2 resistors by mistake with my analogue Amp meter. The meter is now connected in series, inline between the r47's and the output.

I decided to connect up the 50w peltier cooler to see how the PSU performs. Basically one of the 2n3055s seems to be taking most of the load and getting rather hot. I mean I think after about 1 minute it would fail. This is what happened 1st time around. I checked the 7w 1r resistors just incase they were blown, but all gave ok readings.

I then fitted the 10 ohm resistor from the output of the BD139 to the base's of the 2n3055's. Again the transistors got very hot indeed. Two of the transistors don't even get that warm. I connected these on their own if the circuit and they do work, just not sharing the load.

So I think the PSU is just liable to fail due to overload and general heat build up.

The heatsink, which really is just a block of aluminum left over fom a scope build is 15cm L x 5cm W x 1cm D

The input voltage to the circuit from the rectifier is 24V, output variable but maximum load current would be probably 10A, regular use say up to 5A

I like the idea of the current mirrors. I should have a look at this as it may solve the overheat issue. The PSU performs well driving small OC type DC motors and such. It's just the heavier current draws that cause the problems.

Much appreciated again:)

[DrRobin]

Hi,

The 2n3055 emitter resitors you mentioned as 1R not 0.1R, is that correct? I know you measured their resistance, but did you record the values? Differences in these resitors could cause different currents to flow in each transistor, but this should be slight.

The 10 ohm base resistor I suggested should be one for each transistor, commoned at the output of the BD139, not one commoned at the base of the 2n3055s, but that isn't the problem here, something else is wrong.

Can you take voltage reading of the emitters on the 2n3055s under load, that might give me a clue as to why only one is taking the load.

I think we should get this sorted before moving on to current mirrors, they might not help if there is another problem somewhere in the circuit.

I don't think your heatsink is the problem for the uneven current flow, so I think fix that first then I can look at thermal efficency later.

[mindburner]

hello apologies

yes should have been 0.1R. And that's what is fitted in circuit.

I'm in work but wiil measure and record the values with my meter and get back to you.

I will also correct the 10 ohm resistor setup.

I'll also check the output voltage of each 2n3055 under load and report back.

I was first thinking that there was a continuity problem somewhere but all checks out ok. The transistors are all new so hopefully should be fine.

I think I am punching above my weight in trying to build this PSU:) so I totally appreciate the help

PS I also should add that I upped the smoothing from 6000 to 30000uf as I had the caps spare, although this probably won't affect the issue with the transistors

[DrRobin]

Hi,

Not a problem, tonight or when you finish work will be fine. Also at work so just looking at this whilst waiting for my computer to re-build some designs.

Debugging electronic circuits is just the same as anything else, if the basic circuit is good (looks okay) then it's just a case of working through what might be wrong. We have all been there scratching our heads pouring over measurements and can't spot the obvious mistake, same with telescopes, cameras, cars, bike, the whole lot.

You wouldn't believe how long I once spent (went in to weeks) trying to get an rf amp working and I had access to state of the art analysers costing many thousands of pounds, only to find that I hadn't soldered the main transistor in correctly, duh.

Sometimes the best approach is to simplify the situation and the answer usually pops out, like magic.

You could for example, disconnect all the 2n3055s (or at least the base drive) and then test them in circuit one at a time to make sure that they all supply load individually. This would at least tell us that on their own they all work, after that, it has to be some sort of problem with matching transistor characteristics.

Smoothing caps are unlikely to fix the transistor load share issue, but will make the output much smoother.

[mindburner]

hi

I think half my problem is that I try to quite a few things at the one time, usually with reasonable success, but sometimes running into roadblocks like this one.

I teated the PSU with the 50w peltier at 12v. So that's 4.25w I believe.

All four 2n3055's worked individually and produce the 12v output that I selected on the pot. They all get roasting after about 15 seconds and would fail I think if I did not power off.

Each 2n3055 output was at 12v then when the load was applied, rose to 13V

The resistors all measure 0.2 ohm on my peak LCR meter &- 0.2 to -0.4UH inductance

seems that they are all in spec?

cheers

a couple of images of the rats nest:

d

Hi,

Not a problem, tonight or when you finish work will be fine. Also at work so just looking at this whilst waiting for my computer to re-build some designs.

Debugging electronic circuits is just the same as anything else, if the basic circuit is good (looks okay) then it's just a case of working through what might be wrong. We have all been there scratching our heads pouring over measurements and can't spot the obvious mistake, same with telescopes, cameras, cars, bike, the whole lot.

You wouldn't believe how long I once spent (went in to weeks) trying to get an rf amp working and I had access to state of the art analysers costing many thousands of pounds, only to find that I hadn't soldered the main transistor in correctly, duh.

Sometimes the best approach is to simplify the situation and the answer usually pops out, like magic.

You could for example, disconnect all the 2n3055s (or at least the base drive) and then test them in circuit one at a time to make sure that they all supply load individually. This would at least tell us that on their own they all work, after that, it has to be some sort of problem with matching transistor characteristics.

Smoothing caps are unlikely to fix the transistor load share issue, but will make the output much smoother.

[DrRobin]

Hi,

I am like you start about 4 things at once and then never really finish any of them.

Okay, some progress, at least we know that all of the 2n3055s work on their own and they all get hot after a short time.

We also know that the emitter resistors are all around the same value, so that's good as well.

50W Peltier and 12V is around 4.25W, but that is the power dissipated in the Peltier not the transistor. Can you measure the input voltage on the collector of the 2n3055, that way I can work out the power dissipation in the transistors?

Now, the big question is why don't the 2n3055s load share when they are all connected, they should balance it somewhere near even. If I had to guess, I would suggest a wiring problem with the 0.1R emitter resistors, but if you measure the emitter voltages with them all running I might be able to tell.

The 0.1R resistors connect in series with the emitters of the 2n3055s and the common point which is connected to the two 0.47R resistors in parallel and then on to the output. It might be worth you double checking this?

[DrRobin]

Hi,

Just another thought, the emitter resistors are quite well spaced out and then there is cable connecting them to the common point. The cable, although thick (I am guessing it is 1mm2?) has a restistance of about 0.02 ohms per metre, so provided that your lengths are not too uneven, that should be okay.

I am willing to be that the 2n3055 that gets the warmest is closest (in cable length) to the BD139?

1mm2 has a derated current of about 10A, so you are okay, but I might have used 2.5mm2.

[mindburner]

hi

yes the 2n3055 emitters all connect to the resistors via the orange wires 1.2mm I think, you can just about see them in the pictures. The other ends of the resitors are commoned as 2 'pairs' with a lenght of wire between them. One end goes back to the pcb & the LM and the other to the r47's in parallel. The other ends of them go to the amp meter in and the amp meter out to the + output

emitter outputs all seem much the same at 12.7 v when 12v selected on the panel. It's strange

Yes the warmest 2n3055 is closest to the bd139

[DrRobin]

Hi,

It has to be differences in cable length. So I would suggest the following:

Move all four 0.1R emitter resistors together and position them close to the 0.47R resistors. Common one side of the 0.1R resistors, using the resistor legs in to the 0.47R resistors. If they are close together they will all get to the same temperature and so any changes in resistance due to temperature should be minimal.

Wire wound resistors usually have a positive temperature coefficient, so as more current flows it's resistance will increase and this should serve to stabiise the circuit. If they are negative coefficient then changes in temperature will de-stabilise the circuit, but I doubt that it would do so to an extent that you are seeing.

Then connect the other side of the 0.1R resistors to the 2n3055 emitters using the 1mm2 cable, try to make these the same length.

Try to connect the BD139 to the four 2n3055 transistors using the same length of cable and csa, one piece of cable for each, commoned at the BD139. It doesn't need to be 1mm2 as the current is a lot less. You might want to fit 50 or 100 ohm resistors between the BD139 and the base of each 2n3055s but this might not make any difference.

In otherwords, what we are trying to achieve is to balance everything up as best as we can. If it is still very wrong, measure base - emitter voltages of each of the four 2n3055s using a multimeter with probes directly on the back of the 2n3055s and post the results.

[mindburner]

hi

well I have been busy in between putting my son down to sleep. He's over now so managed to get the resistors grouped together, commoned and connected directly to the r47's then to the amp meter then to + out

all the 2n3055 bases are commoned using equal lengths of wire and then using short wire to the BD139

No change, both the left most 2n's get very hot, one more so, same as before

I took a reading using the DVM base to emitter and the 2 left 2n's which don't get hot are at 6v when 5 v selected on the panel, driving the peltier.

The 2 that get hot sit at 5.6v

so I guess there is nearly half a volt there, would that make a difference

thanks again

[DrRobin]

Hi,

Hmm, the voltage difference is significant. A transistor turns on at about 0.6V and is usually fully on by 0.7V base emitter voltage.

The values you gave me, I am assuming that the output is 5V and the two measurements at 5.4V are on the emitter of the two transistors which get hot? The two that don't get hot have an emitter voltage of 6V with the negative probe of the DVM to ground?

If that is correct, then the base - emitter voltage on the two which don't get hot will be much lower, I.E. not driving as hard and hence won't get as hot. However, what I can't understand is if the resistors are all the same value why the emitters of the two which don't get warm are so high.

Just another thought, the transistors should be mounted on insulating washers and plastic spacers round the mounting screws. They then should have a connection to the mounting screw. I can't see the cable coming away from the transistor case or screw, are they connected through the metal back panel or heatsink by any chance? This might well cause a problem.

Finally are all the 2n3055s connected correctly? I know it sounds daft, but at the moment I can't get the 2n3055 voltages to tie up with the heat in two transistors.

[Auntystatic]

What I don't understand is if your just putting 1A through each one, why are 2 of them getting so hot, I've run more than that though them and they barely got warm.

I'm just wondering the same as robin, how are your collectors tied together? I'm also a bit concerned with your mains inlet to the case, the cable will chafe in that hole and there will be a bit of a bang, you need a proper grommet in that hole to protect the cable, make sure the earth is well connected to the case also, unfortunately I'm speaking from an electrocuted frog like experience, boing across the room we fly, you can nick them out of old computer psu's from the output side.

[DrRobin]

Hi,

The relatively low current doesn't really stack up with the quick temperature rise, unless there is a lot of volt drop over the transistors. This would be the case if say the output of the transformer was 30V.

Then 5V output at 1A would generate 25W of heat. This would quickly warm up.

I am going to redraw the cuircuit around the transistors to simplify it and then put on what I know in terms of voltage, if you can fill in the rest please? From what I have seen so far it points to a connection/circuit error, but there might be something else going on.

[mindburner]

Hi,

Hmm, the voltage difference is significant. A transistor turns on at about 0.6V and is usually fully on by 0.7V base emitter voltage.

The values you gave me, I am assuming that the output is 5V and the two measurements at 5.4V are on the emitter of the two transistors which get hot? The two that don't get hot have an emitter voltage of 6V with the negative probe of the DVM to ground?

hi

yes the pots set at 5v

The voltage is base to emitter on each 2n3055, so I had the negative probe on the base of the transistor. The two 2n's with the lower voltage get hot

Just another thought, the transistors should be mounted on insulating washers and plastic spacers round the mounting screws. They then should have a connection to the mounting screw. I can't see the cable coming away from the transistor case or screw, are they connected through the metal back panel or heatsink by any chance? This might well cause a problem.

I checked the continuity again to ensure the heatsink was not connected to anything. All the collectors are tied together and are not shorting with the heatsink. The bases and emitter connections are all heat sleeved and again not shorting. There are mica insulators and inserts fitted. I will maybe refit the transistors just to be sure. I was also wondering if the transistors could be not the best quality as they were ebay specials.

[mindburner]

What I don't understand is if your just putting 1A through each one, why are 2 of them getting so hot, I've run more than that though them and they barely got warm.

I'm just wondering the same as robin, how are your collectors tied together? I'm also a bit concerned with your mains inlet to the case, the cable will chafe in that hole and there will be a bit of a bang, you need a proper grommet in that hole to protect the cable, make sure the earth is well connected to the case also, unfortunately I'm speaking from an electrocuted frog like experience, boing across the room we fly, you can nick them out of old computer psu's from the output side.

hi

The lower screw connects with the case of the transistor and passes through the heatsink, not touching as here is an insulator then a washer. A small eyelet connector attaches with a nut and then wired to the other 3 collectors, then on to the + rail. The upper screw is insulated on both sides.

Good advice in the mains cable. I have a grommet/strain relief on order to fit.

[mindburner]

Hi,

The relatively low current doesn't really stack up with the quick temperature rise, unless there is a lot of volt drop over the transistors. This would be the case if say the output of the transformer was 30V.

Then 5V output at 1A would generate 25W of heat. This would quickly warm up.

I am going to redraw the cuircuit around the transistors to simplify it and then put on what I know in terms of voltage, if you can fill in the rest please? From what I have seen so far it points to a connection/circuit error, but there might be something else going on.

hi

that's great. I will double check all connections and transistor mounting. The tranny is outputting 25v IIRC , 2 12v tied together. I also have some 0r3 resistors on order, so will change them over when they arrive.

[DrRobin]

Hi,

Okay, so far everything looks good with the transistors. They would have to be a hugh mismatch for the results you are getting and since the circuit should be self regulating, I can't see that it is your problem.

I have re-drawn the circuit and added the values on I know, if any are wrong let me know please?

Can you put the system under load and take a set of voltage readings using your DVM please? Connect the negative probe of your DVM to OV or ground and this should be in the same place for all measurements.

Load up the output with a moderate load, not high enough to cause anything to burn out, but high enough to get everything operating, then measure, record and post the following. All measurements should be recorded to as many decimal places as possible (one or two will do).

Voltage on the collectors of all of the 2N3055s. Record as VQ2-C=, VQ3-C=, etc;

Voltage on the base of all of the 2N3055s. Record as VQ2-B=, etc;

Voltage on the emitter of all of the 2N3055s Record as VQ2-E=, etc;

Voltage on the junction of the 0R1 resistors (R1 - R4) and the two 0R47 (R5 & R6). Record as VR5/6=;

Voltage at the output. Record as Vout=.

Let me know by transistor reference which ones get hot. I think you said it was the closest to the BD139, so it will be Q2 & Q3?

If you circuit is slightly different, let me the know the changes.

Based on these results, I should be able to determine what is going wrong. Like you I am keen to get to the bottom of this now, fundermentally it should work and I can't yet understand why it isn't, although the different base voltages you advised before is a clue.

[mindburner]

hi

that's fantastic Robin. I will check and measure all the values and post them in the format requested. Hopefully it will show up the problem. I'm in work at the moment but will get at it as soon as I'm home.

I thought originally that there was a short somewhere but with continuity checking out ok and no actual fail of the PSU, that that side of things were ok.

It's probably something really stupid I have done. It would be nice to get to find the problem. Again thanks for all your time spent with this.

[DrRobin]

Hi,

No problem. It often turns out to be something basic, it's just spotting what it is. Hope I am not giving you too much work to do, but trying to debug something like this remotely isn't easy without all the figures.

I reckon we can get it sorted, it might just chew up a bit of time, still no stargazing at the moment, it's overcast and raining.

[mindburner]

hi

no the figures should be easy enough to get and the best way to solve the imbalance. I was thinking of trying to use regulators like you mentioned before but the PSU is nearly finished I would like to see it through.

Yes no startgazing -BAH. Just managed to get a few shots of the eagle nebula and that was all. Looking forward to seeing more of your m51 as the results you have so far are great, very sharp. I must say I would like one of those Edge's at some point in the future myself.

thanks

Alan