Hi Hans
The second option sounds good to be able to fine tune the phases. Maybe the motor may like 120.3 and 239.9 phases to run the smoothest???
On a different note tried thermistors (5ohm 7amp) and the motor would start and run for a few seconds then slow down and stop. With 2 watt resistors everything is fine. So the only thing I can think of is when thermistor drops down the amps are being hit with current in rush??? Or the motor with the transformer step up is not providing the load the amps want to see???
Here are some numbers. With a 2 ohm resistor load the motor is pulling 2.88 amps at one phase before the transformer after the amp. This is with a 100volts measured across 2 phases. So that is 18 watts per phase wasted heat unless its cold in the room.
Maybe a different thermistor??? Open to ideas here would prefer to lose the heat.
Thanks Tom
The second option sounds good to be able to fine tune the phases. Maybe the motor may like 120.3 and 239.9 phases to run the smoothest???
On a different note tried thermistors (5ohm 7amp) and the motor would start and run for a few seconds then slow down and stop. With 2 watt resistors everything is fine. So the only thing I can think of is when thermistor drops down the amps are being hit with current in rush??? Or the motor with the transformer step up is not providing the load the amps want to see???
Here are some numbers. With a 2 ohm resistor load the motor is pulling 2.88 amps at one phase before the transformer after the amp. This is with a 100volts measured across 2 phases. So that is 18 watts per phase wasted heat unless its cold in the room.
Maybe a different thermistor??? Open to ideas here would prefer to lose the heat.
Thanks Tom
Hi Tom,
There are thermistors with a positive and with a negative temp, did you use the NTC version? And maybe a 7amp version is too big.
I am amazed that your motor is consuming 2.88Amp per phase, but 18Watt per phase would mean a primary trafo voltage of 6.25 Volt ??
If that is the case, why is it so low. The Amps can easily produce twice this voltage, thereby halving the current.
And 100Volt between phases means 58Volt per phase ??
Is this what the motor is asking for?
Hans
Maybe two in paralell would work? That would be 2R5 initial ( say 2R to be safe ). With luck you might loosing less than a 2 R fixed resistor.
I have just measured a SL15 2R509. It gives 1V drop at 2.4A after a few seconds. Doubtless after 5 minutes less. I would guess 0R3 at 3 A. That might half the heat. It seems the only device I have that is close is a good choice on paper. Could be ideal. I thought I had some 5R, alas no.
http://www.ametherm.com/datasheetspdf/SL152R509.pdf
I have just measured a SL15 2R509. It gives 1V drop at 2.4A after a few seconds. Doubtless after 5 minutes less. I would guess 0R3 at 3 A. That might half the heat. It seems the only device I have that is close is a good choice on paper. Could be ideal. I thought I had some 5R, alas no.
http://www.ametherm.com/datasheetspdf/SL152R509.pdf
Hi Hans
The thermistor is a ntc 5d-20 5 ohm / 7 amp. I used it because Ralph used close to that value and it seemed to work.
In regards to to the motor voltage I was told you take the reading from hot to hot so from one lead wire to another. So that would be between 2 transformers not between the output terminals of the same transformer.
To explain it from a different angle I have the 3 step up transformers where one of the output wires goes to one phase of the motor and the other connects to the other 2 wires of the other transformers. So each transformer has a wire to a phase and another to each other.
So instead of measuring across the 2 output wires of each transformer for voltage I was told to measure between the transformers lead to lead. You can see I know as much about 3 phase motors as I do about transistors.
I can increase the voltage with turn of a dial just need to know how to measure it correctly.
Thanks again
Tom
The thermistor is a ntc 5d-20 5 ohm / 7 amp. I used it because Ralph used close to that value and it seemed to work.
In regards to to the motor voltage I was told you take the reading from hot to hot so from one lead wire to another. So that would be between 2 transformers not between the output terminals of the same transformer.
To explain it from a different angle I have the 3 step up transformers where one of the output wires goes to one phase of the motor and the other connects to the other 2 wires of the other transformers. So each transformer has a wire to a phase and another to each other.
So instead of measuring across the 2 output wires of each transformer for voltage I was told to measure between the transformers lead to lead. You can see I know as much about 3 phase motors as I do about transistors.
I can increase the voltage with turn of a dial just need to know how to measure it correctly.
Thanks again
Tom
Hi Guys, the NTC thermistors I am using are 5D-9, i.e 5ohm cold, 3 amp rated.
I did some experimenting with a few different types before settling on this one.
See post #317
I am not surprised that a 7 amp device didn't work properly; the inrush current was probably enough to get the motor started, but the running current was too low to keep the thermistor hot enough to drop its resistance to a working temp.
I did some experimenting with a few different types before settling on this one.
See post #317
I am not surprised that a 7 amp device didn't work properly; the inrush current was probably enough to get the motor started, but the running current was too low to keep the thermistor hot enough to drop its resistance to a working temp.
Last edited:
Ralph,
Have you applied 100Volt per phase to the motor or 100 Volt between the phases?
Hans
As the motor is delta wound it is only possible to feed it with a phase - phase supply. If I generate about 110V rms between the connections on the motor, (see #70), this corresponds to each individual phase-neutral voltage of around 64V rms.
With my transformers, 30VA, 24V ~ 230V I would expect a ratio of around 9.5:1, offload this is not far from what I get, 85Vrms out from a 10Vrms input to the transformer.
Under full load however a 10V input generates a phase-neutral voltage of only 64Vrms, showing the inefficiency of the transformer when being reverse driven.
Hi Ralph
So on your new motor you are hooking up green/yellow/red to phases and purple/brown/blue unconnected for high speed and purple/brown/blue to phases and green/yellow/red unconnected for low speed?
So if I understand you correctly the motor sees the the 110 volts you measured across the phases not the phase neutral of 64 volts. Or to put another way the motors say 117 volt on the side of it so to run it at factory specs we would want to see 117 volts measured from phase to phase.
Have some more ntc's coming 5d-9 and 5d-11 maybe these will get the job done???
Thanks Tom
So on your new motor you are hooking up green/yellow/red to phases and purple/brown/blue unconnected for high speed and purple/brown/blue to phases and green/yellow/red unconnected for low speed?
So if I understand you correctly the motor sees the the 110 volts you measured across the phases not the phase neutral of 64 volts. Or to put another way the motors say 117 volt on the side of it so to run it at factory specs we would want to see 117 volts measured from phase to phase.
Have some more ntc's coming 5d-9 and 5d-11 maybe these will get the job done???
Thanks Tom
Hi Tom, regarding the motor, yes.
The neutral connection is only used to provide a common for the three phases, just like a normal 3 phase generator. That's the beauty of 3 phase; the sum of the currents in the 3 phases always adds to 0.
The correct way to wire the transformers is to connect one secondary wire from each transformer together ( say the 'start' wire) and then insulate it and forget about it. (I recommend a 0.047uF 400V cap between each phase output and the neutral point see #86). You then have 3 wires from the generator to connect to the motor, to change direction just swap any two wires.
Because you're driving the motor with a 3 phase supply the torque is considerably higher, perhaps doubled c/w using a single phase + capacitor setup. This has a secondary advantage that you can run the motor at a significantly lower voltage which reduces the very small residual vibration even further.
With my 'best' motor running at 75~80V its almost impossible to detect with your finger whether the motor is running at all
The neutral connection is only used to provide a common for the three phases, just like a normal 3 phase generator. That's the beauty of 3 phase; the sum of the currents in the 3 phases always adds to 0.
The correct way to wire the transformers is to connect one secondary wire from each transformer together ( say the 'start' wire) and then insulate it and forget about it. (I recommend a 0.047uF 400V cap between each phase output and the neutral point see #86). You then have 3 wires from the generator to connect to the motor, to change direction just swap any two wires.
Because you're driving the motor with a 3 phase supply the torque is considerably higher, perhaps doubled c/w using a single phase + capacitor setup. This has a secondary advantage that you can run the motor at a significantly lower voltage which reduces the very small residual vibration even further.
With my 'best' motor running at 75~80V its almost impossible to detect with your finger whether the motor is running at all
Last edited:
Ralph. That seems a very good result. Sometimes it is worth taking the motor higher as the belt can take a little vibration and the sound might be in some ways better. I built an idler drive turntable with a big Crouzet synchronous ( think I showed it apart ). It sounded fantastic with it's massive torque. Alas rumble was very high. Your motor at 80V might have been the one I needed. In the end I cloned the type of motor typically used. It was always in my mind that rumble might give the sound a sense of drama. In the final version of the turntable - 79 dB weighted rumble and full of drama ( a world record for idler drive ). I am glad to say the vibration isn't the factor we LP lovers worship. However as the LP has rumble it is hard to be certian. Vibrato is the musical term.
This thread shows me how lucky I was with the power supply I used. I did exactly what someone said and used a 150 watt amplifier typically running at 24 VA, this was in 1992. It seems it was tought enough. The amp had no current limiting at all except the output devices. As far as I know none have failed. Using the maths from 100V line 50 amps might be the switch on peak if worst case. The outputs were Hitachi MOS FET's that have a very high Drain Source resistance. It seems the assumption I made at the time was near perfect. I have made similar amps using T03 bipolar Darlington's with 0R33 emitter resistors that were OK. They seemed less happy. The 100V line guys prefer E&1 transformers over toroid.
For fun I have just bought a 100 watt/ 2 ohms module that is current limited to 7.5 amps. It was the worst possible choice. It will be fun to see if it can run that 24 VA load. I will start with the NTC I measuerd yesterday as it looks possible. I might try what the 100V line people recomend. That is 3R9 in paralell with about 3300 uF high ripple. I hate that idea as it is 4 x 3300 uF connected as a non polar type ( AC to a polar cap ). I could use non polar. That might be OK as many are speaker types for high current. How long they would last is anyones guess.
My ideal idea is the two FET's with perhaps 220R and 1000 uF( cap from gates to sources ). If a wafer switch was set up to nearly short the capacitor ( 10 R ? ) I suspect the timing would be OK, this would be between speeds. It is very possible the residual magnetic flux in the transfomer would help switch 33 to 45 RPM or 45 to 33. My conjecture is low for 10 mS and full voltage after 100 mS.
One good idea with shorting the NTC is it cools down. A simple NE555 plus relay and ten second timer seems a good idea. Relaistically this is a great win win solution. The FET has many problems, it needs a high grade floating supply. Being that the switched on gate may take little current a battery might be OK.
This thread shows me how lucky I was with the power supply I used. I did exactly what someone said and used a 150 watt amplifier typically running at 24 VA, this was in 1992. It seems it was tought enough. The amp had no current limiting at all except the output devices. As far as I know none have failed. Using the maths from 100V line 50 amps might be the switch on peak if worst case. The outputs were Hitachi MOS FET's that have a very high Drain Source resistance. It seems the assumption I made at the time was near perfect. I have made similar amps using T03 bipolar Darlington's with 0R33 emitter resistors that were OK. They seemed less happy. The 100V line guys prefer E&1 transformers over toroid.
For fun I have just bought a 100 watt/ 2 ohms module that is current limited to 7.5 amps. It was the worst possible choice. It will be fun to see if it can run that 24 VA load. I will start with the NTC I measuerd yesterday as it looks possible. I might try what the 100V line people recomend. That is 3R9 in paralell with about 3300 uF high ripple. I hate that idea as it is 4 x 3300 uF connected as a non polar type ( AC to a polar cap ). I could use non polar. That might be OK as many are speaker types for high current. How long they would last is anyones guess.
My ideal idea is the two FET's with perhaps 220R and 1000 uF( cap from gates to sources ). If a wafer switch was set up to nearly short the capacitor ( 10 R ? ) I suspect the timing would be OK, this would be between speeds. It is very possible the residual magnetic flux in the transfomer would help switch 33 to 45 RPM or 45 to 33. My conjecture is low for 10 mS and full voltage after 100 mS.
One good idea with shorting the NTC is it cools down. A simple NE555 plus relay and ten second timer seems a good idea. Relaistically this is a great win win solution. The FET has many problems, it needs a high grade floating supply. Being that the switched on gate may take little current a battery might be OK.
I thought you might like to see my minimal parts FET relay with NTC look to the output. It's crude and more interestingly it works. I had these FET already doing something else, not a prefered type. I was very surprised how long it took to work as I had exspected it to jump on too fast. I have not put it on a data logger to see the real curve. It isn't too expensive to try if wanting something to play with. 12R is what came to hand and not a special value. It seems to be ideal. I don't think a battery can be bettered. A rechargable might suit best. A small lead type even.
Other than a metal contact relay this loss looks the least possible. Some FET will comfortably out do IRF530. Some cost about the same. STP30NF10 perhaps?
Please don't take it that I totally believe this is the best solution. I think it is worth a look. The curve might not suit or vary too much between samples ( I didn't find this ). The capacitor was standard grade just to be sure nothing fancy required. A Panasonic FC ( 470 uF 25 V ) would be nice and cheap. 10K was choosen so as not to be too high nor low. If you see any no-no's please say. I think the 12 R is OK ( choose good ripple grade, >1 amp if possible). It could be that the curve is still too fast. If so 22K could help. Roughly speaking 4 seconds before the meter said we have power. I just hope the vital bit is slow enough. One would guess 100 mS should be fine. That is 2R down to 0R2. If I get a moment I will data log it. The 9V is not shown switched for simplicity. I just used the battery clip terminals for this test. It should be that no heatsinking is required.
I received the prototype PCBs on Tuesday this week and built up one of the generators. Other than a few component changes, everything works as expected. The distortion improved a little from the breadboard version at ~0.5% (-46dB). The second harmonic is the most dominant.
On Wednesday, I completed the user interface for the frequency changing portion of the operating system. The phases are still fixed at 90/120/240 deg. I should have the user interface for the phase adjustments done by the weekend.
A preliminary video of the generator operation can be seen Here.
On Wednesday, I completed the user interface for the frequency changing portion of the operating system. The phases are still fixed at 90/120/240 deg. I should have the user interface for the phase adjustments done by the weekend.
A preliminary video of the generator operation can be seen Here.
Attachments
Last edited:
To reiterate, I have no intention of selling finished controllers or parts. I can "share" the PCB design through OshPark (board fabricator in Oregon) so that someone can click on a link and buy 3 pcs through them for $43.80 including shipping (~$14.60/pcb).
The parts kit can be purchased through Mouser electronics via a "shared" shopping cart that I created. Right now, the kit is $33 including the uP which is un-programmed. I was seeking someone that would buy and stock a number of uP's, so that I could program a group of them (at N/C), lock the security bits so my IP can't be pirated, and send them back for distribution/sale to DIYers.
All of the parts are through hole (the uP which is PLCC is socketed) and it took me less than an hour to build one, but it does require some soldering skill.
Some additional specs:
All 3 outputs are DC coupled from the op-amps, 5VPP and centered at 2.5VDC.
Freq range is limited to 40.00-70.00Hz for 33 RPM and 60.00-90.00Hz for 45RPM.
Freq resolution is 0.01Hz, crystal controlled (100 PPM) using a 20bit DDS core implemented in software. D to A conversion is done using 8 bit PWM running at 18kHz.
The 3 phases will be adjustable in ~1.5 deg steps, ±15 deg (±10 steps).
When the generator first powers up, the EEPROM will be blank; the uP detects this and immediately enters factory default mode to load all of the default parameters. JP1 determines if the default parameters are 50Hz/67.5Hz or 60Hz/81Hz. Factory defaults can be restored at any time by pressing and holding both the UP/DN buttons while applying power. JP1 is only looked at during factory default mode; shorting/removing it during normal operation has no effect.
Phase calibration mode is entered by holding the UP button while exiting standby mode. There are 6 separate phase adjustments: 90/120/240 for each speed (33/45).
The parts kit can be purchased through Mouser electronics via a "shared" shopping cart that I created. Right now, the kit is $33 including the uP which is un-programmed. I was seeking someone that would buy and stock a number of uP's, so that I could program a group of them (at N/C), lock the security bits so my IP can't be pirated, and send them back for distribution/sale to DIYers.
All of the parts are through hole (the uP which is PLCC is socketed) and it took me less than an hour to build one, but it does require some soldering skill.
Some additional specs:
All 3 outputs are DC coupled from the op-amps, 5VPP and centered at 2.5VDC.
Freq range is limited to 40.00-70.00Hz for 33 RPM and 60.00-90.00Hz for 45RPM.
Freq resolution is 0.01Hz, crystal controlled (100 PPM) using a 20bit DDS core implemented in software. D to A conversion is done using 8 bit PWM running at 18kHz.
The 3 phases will be adjustable in ~1.5 deg steps, ±15 deg (±10 steps).
When the generator first powers up, the EEPROM will be blank; the uP detects this and immediately enters factory default mode to load all of the default parameters. JP1 determines if the default parameters are 50Hz/67.5Hz or 60Hz/81Hz. Factory defaults can be restored at any time by pressing and holding both the UP/DN buttons while applying power. JP1 is only looked at during factory default mode; shorting/removing it during normal operation has no effect.
Phase calibration mode is entered by holding the UP button while exiting standby mode. There are 6 separate phase adjustments: 90/120/240 for each speed (33/45).
Last edited:
Hi Guys
Tried the new thermistors ntc 9 and ntc 11 runs for few seconds amps kick out. Also tried parallel no luck. Will work fine with as low as 1.7 ohm resistor but no luck with thermistor.
Could it be the thermistors are acting faster than the impedance can ramp up with this big motor. Maybe the large 50 via transformers need more time to ramp up. Running the transformers parallel both sides for a 12 to 120 step up.
I have some timed relays coming you can program from .01 seconds to hours. So if the motor is showing the amps at least 2 ohms I would have to think they will work. Not happy with 56 watts of heat for no reason.
Open to ideas why the thermistors are not working. Ralph is having good luck but not sure which motor he is using with them.
Thanks Tom
Tried the new thermistors ntc 9 and ntc 11 runs for few seconds amps kick out. Also tried parallel no luck. Will work fine with as low as 1.7 ohm resistor but no luck with thermistor.
Could it be the thermistors are acting faster than the impedance can ramp up with this big motor. Maybe the large 50 via transformers need more time to ramp up. Running the transformers parallel both sides for a 12 to 120 step up.
I have some timed relays coming you can program from .01 seconds to hours. So if the motor is showing the amps at least 2 ohms I would have to think they will work. Not happy with 56 watts of heat for no reason.
Open to ideas why the thermistors are not working. Ralph is having good luck but not sure which motor he is using with them.
Thanks Tom
HI Tom, I'm currently using the ntc9 thermistors, but two other differences.
I'm using a 30VA transformer, and the windings are in series, 24-230v, so the impedances are considerably higher. My setup works with all my motors.
I am in the process of trying some 60VA transformers, I'll let you know how I get on with these.
I'm using a 30VA transformer, and the windings are in series, 24-230v, so the impedances are considerably higher. My setup works with all my motors.
I am in the process of trying some 60VA transformers, I'll let you know how I get on with these.
Graph of FET's as current limiter ( gate voltage, 5.1V = 1R ). When hand logging the resistance it looked too fussy. After logging it this way I find it " might " have a 0.8 second window if using 20K charging cap. If you look at this real graph of the real devices anywhere on the charging curve will do within reason. The thing that controls the rate is V battery and the charge resistor. This is a very cheap cap so shows nothing fancy required. The discharge has been 100% relaible ( 30 mS with about 0.75 A initial current, CR = 5.7 mS ). To get the best possible graph to work with I drew in the missing bit for you to see. 7 seconds seems about the time until low resistance. CR = 9.4 seconds ( would be 6.1 V ) and 5CR about less than 1 minute ( 1CR is enough ). Looks about right I think. This might be 17 times lower loss with identical distortion whilst using cheap parts ( static sensetive ). Vibration might be very slightly less than with a 1R7 resisitor. I will use this for my version. Hope you like it. I hope to prove a NTC a better choice. The time between lines is one second. The battery is 10V more or less.
A big Thank You Pyramid, this is really exciting.
In case it got lost, I'll volunteer to purchase the uP, send them to Pyramid for bulk programming, etc, once we're at that point. I may be able to do the pcb portion too.
Pyramid - are there mounting holes on your pcb? I didnt see them.
Would it be possible to release a BOM? personally I refuse to use Mouser.
The Phoenix Engineering tachometer can be used as a standalone piece of equipment, right?