I'm not sure what effect the 'damping factor' will have on an amplifier driving a transformer at a constant level and frequency (I hope).
The only advantage I can see is to reduce the heat generated in the enclosure, but even here it's only in the order of a couple of watts.
I have slightly more trouble with shielding the NTC thermistors from the airflow produced by the small fan that I fitted to keep the output transformers adequately cooled, I may have to wrap the thermistors in small glass fibre shrouds, which has the further advantage of letting them warm up more and dropping their resistance even more.
In theory like a stiff car suspension damping does what it big change. As people like a full picture of a subject it is a reasonable question.
Hans. If you look closely the FET's are doing opposite half cycles. If not the FET will be reversed biased and conducting. However you may have a point that only the FET's get involved, it looks too perfect to involve the diodes much. Being 13.5 VAC it would show. This is the conventional wisdom below. Anyway. I wanted to show an unlikely idea that can work and cost peanuts.
Wiki quote.
"An SSR based on a single MOSFET, or multiple MOSFETs in a paralleled array, can work well for DC loads. MOSFETs have an inherent substrate diode that conducts in the reverse direction, so a single MOSFET cannot block current in both directions. For AC (bi-directional) operation two MOSFETs are arranged back-to-back with their source pins tied together. Their drain pins are connected to either side of the output. The substrate diodes are alternately reverse biased to block current when the relay is off ( that's the key point ). When the relay is on, the common source is always riding on the instantaneous signal level and both gates are biased positive relative to the source by the photo-diode ( or whatever ).
It is common to provide access to the common source so that multiple MOSFETs can be wired in parallel if switching a DC load. Usually a network is provided to speed the turn-off of the MOSFET when the control input is removed.
Hans. If you look closely the FET's are doing opposite half cycles. If not the FET will be reversed biased and conducting. However you may have a point that only the FET's get involved, it looks too perfect to involve the diodes much. Being 13.5 VAC it would show. This is the conventional wisdom below. Anyway. I wanted to show an unlikely idea that can work and cost peanuts.
Wiki quote.
"An SSR based on a single MOSFET, or multiple MOSFETs in a paralleled array, can work well for DC loads. MOSFETs have an inherent substrate diode that conducts in the reverse direction, so a single MOSFET cannot block current in both directions. For AC (bi-directional) operation two MOSFETs are arranged back-to-back with their source pins tied together. Their drain pins are connected to either side of the output. The substrate diodes are alternately reverse biased to block current when the relay is off ( that's the key point ). When the relay is on, the common source is always riding on the instantaneous signal level and both gates are biased positive relative to the source by the photo-diode ( or whatever ).
It is common to provide access to the common source so that multiple MOSFETs can be wired in parallel if switching a DC load. Usually a network is provided to speed the turn-off of the MOSFET when the control input is removed.
One critisism of my FET circuit is I didn't include a discharge resistor gates to sources. On my real version 1K input and 10K gates/sources. If a useful switch on effect was to be had this simple arrangement looks unlikely to work. Some circuits seem to show it might. Of most importance is low distortion and can be very fast. I agree the NTC device is so simple. I use them a lot. I ran the circuit shown up to 115 VAC without problems before buying T0247 650 VDC devices. You have to run the gates as hard as you can to ensure the FET is fully saturated. If not they will get very hot. Look to see maximum gate voltage and typical fully on maximum voltages. Usually that gives scope to turn them very hard on.
Giving the reversed biased MOSFET some thought. It can conduct both ways if the gate is on which must be why no obvious distortion is seen and heat so low ( Thanks Hans for making me think it through ). It brings into question when the diode most often works. When in a class AB amp I guess it must when needed. This answers a question I had struggled with, when does this diode come in ( often shown as a zener ). I guess if the resistance of the FET is circa 0R5 and we have 2 amps the diode should take over. That's interesting. Time to buy some more big FET's to test the idea. The IRF640 I used first most likely could offer more than 2 amps without distortion. From the volts dropped it was book specification. I did try some Schotkey diodes and still had identical results. The drain source resistance must be very low when IRF640.
BTW. If my damping factor caution is true the FET if used must be a very low resistance type. As luck has it not difficult at lower voltages. The more I look at this very reasonable precaution I would use a fixed resistor and old fashioned relay with NE555 timer.
Giving the reversed biased MOSFET some thought. It can conduct both ways if the gate is on which must be why no obvious distortion is seen and heat so low ( Thanks Hans for making me think it through ). It brings into question when the diode most often works. When in a class AB amp I guess it must when needed. This answers a question I had struggled with, when does this diode come in ( often shown as a zener ). I guess if the resistance of the FET is circa 0R5 and we have 2 amps the diode should take over. That's interesting. Time to buy some more big FET's to test the idea. The IRF640 I used first most likely could offer more than 2 amps without distortion. From the volts dropped it was book specification. I did try some Schotkey diodes and still had identical results. The drain source resistance must be very low when IRF640.
BTW. If my damping factor caution is true the FET if used must be a very low resistance type. As luck has it not difficult at lower voltages. The more I look at this very reasonable precaution I would use a fixed resistor and old fashioned relay with NE555 timer.
Last edited:
Hi Guys
Got the 3 phase motor working with the old generator. So I built a new Hans 3 phase splitter lower part of post 151. I put in +15 volts to pin 8 and -15 volts to pin 4. Also a .1 cap from pin 8 to ground as and a .1 cap from pin 4 to ground. The rest just like the diagram.
I am having a issue with the output of the splitter. The 0 degree is at .381 volts the 120 degree is at .299 volts and the 240 degree is at .299 volts.
I can adjust at Ralph's voltage divider/low pass filter before the amps but do not know what that does to degrees of the phases??? Any ideas???
Thanks Tom
Got the 3 phase motor working with the old generator. So I built a new Hans 3 phase splitter lower part of post 151. I put in +15 volts to pin 8 and -15 volts to pin 4. Also a .1 cap from pin 8 to ground as and a .1 cap from pin 4 to ground. The rest just like the diagram.
I am having a issue with the output of the splitter. The 0 degree is at .381 volts the 120 degree is at .299 volts and the 240 degree is at .299 volts.
I can adjust at Ralph's voltage divider/low pass filter before the amps but do not know what that does to degrees of the phases??? Any ideas???
Thanks Tom
Since the 240 degrees has exactly the same voltage as the 120 degrees, the error must be with the first op amp making the 120 degrees.
There must be some bad connection or short circuit.
What is the DC offset at the output of both op amps when you replace the 60 Hz at the input by a short circuit to ground.
And what is the output voltage of both op amps when you apply 1.5 V DC (with a battery) to the input.
Output of both op amps should be exactly the same as the input voltage.
Concentrate on the first op amp, there is the error to be found.
Hans
Can you check that resistors R3 and R5 are the right way round, i.e not swapped? This is the only way I can find that corresponds with the results you are getting.
pps, no sign of any attachments to your email I'm afraid. I think I've worked out the motor windings from first principles, blue,brown and violet for one speed, yellow green and red for the other.
pps, no sign of any attachments to your email I'm afraid. I think I've worked out the motor windings from first principles, blue,brown and violet for one speed, yellow green and red for the other.
Hans/Ralph
I do not know how to check DC offset. Do not know what it is. In regards to 1.5 battery voltage op amp check do you bypass the 13k9 resistor and put the voltage to pin3 and check output at pin 1 and same to pin 5 output at pin 7. I am assuming this is powered up and no 60 hz input signal.
Texas has their diagram with the + at the top of triangle and - on bottom. Hans has it the opposite. So I just hooked up according to the +/- signs. So the 10k is going to -2 and-6 input and 13k9 is going to +3 and +5 inputs.
Thanks Tom
PS Ralph will try another angle on attachment.
I do not know how to check DC offset. Do not know what it is. In regards to 1.5 battery voltage op amp check do you bypass the 13k9 resistor and put the voltage to pin3 and check output at pin 1 and same to pin 5 output at pin 7. I am assuming this is powered up and no 60 hz input signal.
Texas has their diagram with the + at the top of triangle and - on bottom. Hans has it the opposite. So I just hooked up according to the +/- signs. So the 10k is going to -2 and-6 input and 13k9 is going to +3 and +5 inputs.
Thanks Tom
PS Ralph will try another angle on attachment.
Hi Tom,
I mentioned to remove the 60Hz connection and to apply a short circuit instead, but leave everything else intact.
Offset means, how many millivolts do you measure between op amp out and ground with this short circuit, so this is an offset with respect to ground.
With the 1.5 Volt leave everything in place, remove the short circuit and apply this voltage to the input. Then measure again what voltage you measure at the output of the op amp.
If you prefer, you could also send a sketched diagram with the exact information on pins, power supply and ground connections.
Hans
How's it going now? Are you pleased with NTC thermistors? I read up on Ti site about 100 Volt Line and class D, a very similar problem. Funny thing is they don't really give an answer except to say the signal should be ramped. I suspect the NTC solution is by far the best now I have considered all the twists and turns. I still like the relay in addition.
I met the guy who designed the Revox ( so I was told ). A really nice gentleman ( Think Arnold Ridley ). He was with a Mr Stein who repairs Quad speakers. I have noticed the really great people of Audio tend to be very modest. It can be a very powerful surprise when one learns what they designed.
The NTC thermistors are working almost perfectly, with no more than 0.5V drop when hot.
The only very small downside occurs if you switch the amps off and back again within a few seconds, when the thermistors are still hot, the amps can lockup. Even then the amps reset themselves a few seconds later and the motor spins up normally.
I'm happy with the setup as it now is; just looking into programming my AD9833 chips, but that may take some time
.
The only very small downside occurs if you switch the amps off and back again within a few seconds, when the thermistors are still hot, the amps can lockup. Even then the amps reset themselves a few seconds later and the motor spins up normally.
I'm happy with the setup as it now is; just looking into programming my AD9833 chips, but that may take some time
.This was the only idea I had that looked simple enough. If the lamp can hold down the FET's long enough it could work. The gate voltage would need to go as high as possible within the spec permitted to allow for many factors. The likely problem is the lamp might not do all the things required. Best say the NTC is a good idea. Thanks for doing the finding Ralph.
Hi Hans
Connected 60hz lead to ground and measured DC from pin 1 to ground 6.7 MV. Then took those leads and put 1.569dcv on input with 1.563 at pin 1.
Yesterday I tried different op amp no change. The generator was putting out almost 1 volt DC to amps so I put a 1 uf cap in series which got rid of the DC but no change in the voltage drop. I also resoldered the joints all around op amp again no change.
Put in picture of schematic I built from.
Thanks for all the help
Tom
Connected 60hz lead to ground and measured DC from pin 1 to ground 6.7 MV. Then took those leads and put 1.569dcv on input with 1.563 at pin 1.
Yesterday I tried different op amp no change. The generator was putting out almost 1 volt DC to amps so I put a 1 uf cap in series which got rid of the DC but no change in the voltage drop. I also resoldered the joints all around op amp again no change.
Put in picture of schematic I built from.
Thanks for all the help
Tom
Attachments
Hi Tom,
All DC measurements are correct, so most connections are OK.
Interchanging 10k and 13k9 does also not bring the ratio from .381 to .299 volt. So the only thing that remains is that the 330nF cap is defect or not connected correctly.
Take this cap out from pin3 to Gnd and put again 60Hz at the input with no cap.
If all 3 phases now have the same amplitude, it was definitely the cap.
Awaiting your results,
Hans
Hi Hans
Hans you are one sharp cookie. Cut the ground on the cap in question and voltages are now close. Measured cap and it measures OK go figure. Of So I put in a new bigger size film 330 nf (343 to be exact). And we have success.
Did solder joint twice and cap measures fine out of circuit. It was only a 50v tant cap??? Well it is going in the trash!!!
Should I run a input cap to get rid of the one volt dc coming out of generator? Is 1 uf a good value there?
What can I say but thanks again Hans
Tom
Hans you are one sharp cookie. Cut the ground on the cap in question and voltages are now close. Measured cap and it measures OK go figure. Of So I put in a new bigger size film 330 nf (343 to be exact). And we have success.
Did solder joint twice and cap measures fine out of circuit. It was only a 50v tant cap??? Well it is going in the trash!!!
Should I run a input cap to get rid of the one volt dc coming out of generator? Is 1 uf a good value there?
What can I say but thanks again Hans
Tom
Hi Tom,
You could insert a cap if you like, but there is no direct need to do that, unless the 60Hz has many harmonics that should be filtered.
In this simple setup, the relation between the 3 phases always remain intact, no matter what you place in between after the 60Hz generator.
Now that everything seems to be working, you will have to fine adjust the 120 and 240 degrees.
To achieve this, there are two ways.
First is when you know the exact value of the caps, to newly calculate the right value for the 13k9 resistor.
The product of both should be 4594e-6.
So for your 343nF the resistor should be 13K39 (=365K par to 13k9)and so on for the second 13k9 resistor.
Second way is to replace the 13k9 resistors by 11k4 resistors in series with 5K multiturn trimmer pots, giving you a +/-18% adjustment range.
With the three 10k resistors connected to the 3 phases, by turning the pots you should reduce this check signal to the absolute minimum.
Hans