Phase Perfect dpc-a10 - Page 5
Close
Login to Your Account
Page 5 of 6 FirstFirst ... 3456 LastLast
Results 81 to 100 of 110
  1. #81
    Join Date
    May 2019
    Country
    UNITED STATES
    State/Province
    Pennsylvania
    Posts
    195
    Post Thanks / Like
    Likes (Given)
    39
    Likes (Received)
    46

    Default

    I was under the impression that a VFD needed to start and stop the load and could not take the inrush of startup or the spikes associated with disconnecting a load. I used to have two 75hp ABB VFDs I threw out years ago thinking I could never use them (there were a dozen of them used for temporary cooling towers for 2 years and then they all went in the trash along with the rest of the temporary cooling tower setup)..... I often have a couple sources for getting large industrial gear dirt cheap.

  2. #82
    Join Date
    May 2019
    Country
    UNITED STATES
    State/Province
    Pennsylvania
    Posts
    195
    Post Thanks / Like
    Likes (Given)
    39
    Likes (Received)
    46

    Default

    I robbed one of the new current sensors for my old board and the unit now cycles on normally and the contactor closes and the green light is on, but L1-L2 242 volts but L1-L3 and L2-L3 are 120 volts. Not sure if it needs a load on it or if it has something to do with the board being remotely mounted and no wires going through the current sensors but we are headed in the right direction......I think

  3. #83
    Join Date
    Aug 2014
    Location
    bainbridge island
    Posts
    1,129
    Post Thanks / Like
    Likes (Given)
    213
    Likes (Received)
    267

    Default

    Quote Originally Posted by DavidScott View Post
    Yeah, VERY! As backup and I now have 2 cnc mills but can only run one at a time because my phase converter is too small. Do you want it?
    Take it, i have 3 phase at work.

  4. Likes DavidScott liked this post
  5. #84
    Join Date
    Jun 2001
    Location
    St Louis
    Posts
    17,834
    Post Thanks / Like
    Likes (Given)
    1780
    Likes (Received)
    2966

    Default

    Quote Originally Posted by Hardplates View Post
    I was under the impression that a VFD needed to start and stop the load and could not take the inrush of startup or the spikes associated with disconnecting a load. I used to have two 75hp ABB VFDs I threw out years ago thinking I could never use them (there were a dozen of them used for temporary cooling towers for 2 years and then they all went in the trash along with the rest of the temporary cooling tower setup)..... I often have a couple sources for getting large industrial gear dirt cheap.
    Well, the usual VFD has a 150% overload for an extended time, perhaps a minute or even two. The longer time ones may have a 200% overload rating that is a few seconds. If properly designed, that is OK at any operational temperature.

    The inrush is a problem if the VFD is "matched" to the load. a 10A motor and a VFD that can supply 10A (usually a bit more) indefinitely, for instance. But, if you power that 10A motor with a 30A VFD that gas a short term 200% overload, you "can" start a motor which takes 60A at start. I'd not cut it that fine, but it nominally works, at least.

    So if the inrush is within the VFD's current capability, then it is not an issue. I had to make a VFD that was to start and run a small military oil burner based heater unit. The motor took 2.2A, but the surge was over 12A, so it had to supply that on a short term basis. It was really just larger than absolutely required IGBTs, suitable gate drive, and some software, because the surge was so short term that the bus voltage supply did not care.

    So the usual PP would have to supply enough surge to start the largest motor it can drive. And, possibly the highest start current version of that size motor, since they can vary 2:1 in start current, possibly more. Most can be at least 5x FLA, so the most may be 10x or maybe 12x FLA.

    It just takes bigger IGBTs and suitable gate drive, programming, along with the basic capability of supplying that current through the wires, etc, without too much voltage sag.

    Use of a pp, or a large (oversized) VFD, either can do the job. One is a lot less trouble to set up and install, you just throw money at it.

  6. #85
    Join Date
    May 2019
    Country
    UNITED STATES
    State/Province
    Pennsylvania
    Posts
    195
    Post Thanks / Like
    Likes (Given)
    39
    Likes (Received)
    46

    Default

    Quote Originally Posted by JST View Post
    Well, the usual VFD has a 150% overload for an extended time, perhaps a minute or even two. The longer time ones may have a 200% overload rating that is a few seconds. If properly designed, that is OK at any operational temperature.

    The inrush is a problem if the VFD is "matched" to the load. a 10A motor and a VFD that can supply 10A (usually a bit more) indefinitely, for instance. But, if you power that 10A motor with a 30A VFD that gas a short term 200% overload, you "can" start a motor which takes 60A at start. I'd not cut it that fine, but it nominally works, at least.

    So if the inrush is within the VFD's current capability, then it is not an issue. I had to make a VFD that was to start and run a small military oil burner based heater unit. The motor took 2.2A, but the surge was over 12A, so it had to supply that on a short term basis. It was really just larger than absolutely required IGBTs, suitable gate drive, and some software, because the surge was so short term that the bus voltage supply did not care.

    So the usual PP would have to supply enough surge to start the largest motor it can drive. And, possibly the highest start current version of that size motor, since they can vary 2:1 in start current, possibly more. Most can be at least 5x FLA, so the most may be 10x or maybe 12x FLA.

    It just takes bigger IGBTs and suitable gate drive, programming, along with the basic capability of supplying that current through the wires, etc, without too much voltage sag.

    Use of a pp, or a large (oversized) VFD, either can do the job. One is a lot less trouble to set up and install, you just throw money at it.
    Interesting and good to know. I've never played around with VFD's outside their intended purpose but now that you mention it that may be a fun project to play around with.

    On another note there still appears to be an intermittent mechanical fault with my original board as sometime the LED's don't light up and the contactor doesn't close and if I apply a twisting or tapping force to the board it cycles on normally and closes the contactor but only has 120 volt between L1-L3 L2-L3. It operates the same if I disconnect the Z5 gate and does not throw the blinking red light. I am unsure why it doesn't recognize the Z5 gate being disconnected but I believe I will head in the direction of tracing out the gate driver circuit and look for a signal.

    Any ideas? Please share. Has anyone ever had a failure of a phase perfect that still had a green LED and closed contactor? Of the limited number of people I have spoken with, when they experienced a failure the unit threw some kind of "code" with the LED's.

  7. #86
    Join Date
    Mar 2005
    Country
    UNITED STATES
    State/Province
    Georgia
    Posts
    134
    Post Thanks / Like
    Likes (Given)
    60
    Likes (Received)
    56

    Default

    Quote Originally Posted by Hardplates View Post
    I robbed one of the new current sensors for my old board and the unit now cycles on normally and the contactor closes and the green light is on, but L1-L2 242 volts but L1-L3 and L2-L3 are 120 volts. Not sure if it needs a load on it or if it has something to do with the board being remotely mounted and no wires going through the current sensors but we are headed in the right direction......I think
    Much progress has been made, right direction for sure. I still think it is best to pursue fixing this unexpected low voltage output before the mechanical fault, because it may turn out that the mechanical fault is near irreparable due to some internal problem with the PC board itself, and more twisting and tapping may only render it useless for further learning about repairing these units.
    I think there is at least some possibility that the PP does not put full voltage on the manufactured leg until it senses what sort of load is on it. In other words, it may need to experience a load of some sort before you can measure this voltage. If you have nothing else, put two 120V 100W incandescent lamps in series between this and one of the other legs, see if that makes a difference.

  8. #87
    Join Date
    May 2019
    Country
    UNITED STATES
    State/Province
    Pennsylvania
    Posts
    195
    Post Thanks / Like
    Likes (Given)
    39
    Likes (Received)
    46

    Default

    Quote Originally Posted by RustySparks View Post
    Much progress has been made, right direction for sure. I still think it is best to pursue fixing this unexpected low voltage output before the mechanical fault, because it may turn out that the mechanical fault is near irreparable due to some internal problem with the PC board itself, and more twisting and tapping may only render it useless for further learning about repairing these units.
    I think there is at least some possibility that the PP does not put full voltage on the manufactured leg until it senses what sort of load is on it. In other words, it may need to experience a load of some sort before you can measure this voltage. If you have nothing else, put two 120V 100W incandescent lamps in series between this and one of the other legs, see if that makes a difference.
    That exact though crossed my mind and because of it I used lowZ to measure the voltage, but perhaps that is not enough load to get it to start up. I will try your idea of light bulbs as a load.

    YouTube

    In this video the guy has voltage, but who knows what is on the other end of the lines or if the lines themselves are enough of a initial load. My better judgment tells me it won't be so simple but you are correct and the theory needs to be tested before moving forward in search of other faults.

  9. #88
    Join Date
    May 2019
    Country
    UNITED STATES
    State/Province
    Pennsylvania
    Posts
    195
    Post Thanks / Like
    Likes (Given)
    39
    Likes (Received)
    46

    Default

    I'm working with a 2 channel scope so there are two photos but both are Triggered off channel 1 (blue) which is T3.

    In one photo channel 2 (red) is T1 and is the other photo channel 2 (red) is T2 but is always being triggered off of channel 1 (blue) T3

    There is a 100 watt load between T2 and T3.


    T3 appears to be at the correct phase angle but only about 20 volts. Its trying and thinks its working......but obviously is only kinda working.....any ideas?
    Attached Thumbnails Attached Thumbnails t3-t1.jpg   t3-t2.jpg  

  10. #89
    Join Date
    Aug 2014
    Location
    bainbridge island
    Posts
    1,129
    Post Thanks / Like
    Likes (Given)
    213
    Likes (Received)
    267

    Default

    If the output voltage sensing network is damaged it may think it is delivering full voltage.

    Does it have a transformer and then a resistive voltage divider in front of an opamp or an Adc?

  11. #90
    Join Date
    Mar 2005
    Country
    UNITED STATES
    State/Province
    Georgia
    Posts
    134
    Post Thanks / Like
    Likes (Given)
    60
    Likes (Received)
    56

    Default

    So, no load 120 volts, light bulb load 20 volts? Normal looking waveform and phase angle? High resistance connection in the path to the terminal where you are measuring MIGHT be the problem. I assume there is a contractor that switches on the output when system thinks all is well. Any chances it has a burned contact on the generated leg making poor connection? If there was a problem with damaged output voltage sensing network, I would think the voltage would be way too high, or at least would be the same 120VAC with or without your load.

  12. #91
    Join Date
    May 2019
    Country
    UNITED STATES
    State/Province
    Pennsylvania
    Posts
    195
    Post Thanks / Like
    Likes (Given)
    39
    Likes (Received)
    46

    Default

    Quote Originally Posted by RustySparks View Post
    So, no load 120 volts, light bulb load 20 volts? Normal looking waveform and phase angle? High resistance connection in the path to the terminal where you are measuring MIGHT be the problem. I assume there is a contractor that switches on the output when system thinks all is well. Any chances it has a burned contact on the generated leg making poor connection? If there was a problem with damaged output voltage sensing network, I would think the voltage would be way too high, or at least would be the same 120VAC with or without your load.
    The 120 volts was phase to phase. The 20 volts is to ground. The contactor and such seems ok, but I had a loose connection and when I did the IGBT was outputting 100 volts to ground. Maybe there is something dragging the driver down, I must be missing something simple.

    Should be easier to troubleshoot tomorrow when I get back with the other unit. I also have to work on my Hardinge superslant turning center, and do school work for college and oh yeah, make some money to pay the bills. Oh and maybe pick up a Star JNC-10 swiss style lathe.

  13. Likes RustySparks liked this post
  14. #92
    Join Date
    May 2019
    Country
    UNITED STATES
    State/Province
    Pennsylvania
    Posts
    195
    Post Thanks / Like
    Likes (Given)
    39
    Likes (Received)
    46

    Default

    Much thanks to thermite cause now I have a working unit to compare against.

    When referring to noise I am speaking of noise in the waveform and not noise you hear.

    There appears to be some phase noise in relation to ground. It does not show from phase to phase. I replaced the AC caps, although the old ones seem fine with an ESR of .6-.9ohms. At this point I am not sure if this is a problem or if it is inherent to these units. My friends 30hp unit has similar noise although I only checked to ground and not phase to phase at the time. I would imagine it would be fine as it is to ground and I am not running a neutral since all 3 phase loads I have that also use single phase do so with a transformer and floating neutral.

    If anyone has any ideas I am all ears, or it would be great if someone with a PP and a scope could check the waveform on their unit. At this point I am unsure if this is something that needs to be fixed or if they were like this when new.
    Attached Thumbnails Attached Thumbnails pp2.jpg   pp1.jpg  

  15. #93
    Join Date
    May 2019
    Country
    UNITED STATES
    State/Province
    Pennsylvania
    Posts
    195
    Post Thanks / Like
    Likes (Given)
    39
    Likes (Received)
    46

    Default

    This is the "no so" DC bus. Maybe filter caps or there is a cap hooked to the chassis....could it be the issue?..... More investigation to follow
    Attached Thumbnails Attached Thumbnails ppdc1.jpg  

  16. #94
    Join Date
    May 2019
    Country
    UNITED STATES
    State/Province
    Pennsylvania
    Posts
    195
    Post Thanks / Like
    Likes (Given)
    39
    Likes (Received)
    46

    Default

    This is very far from what I would call a "perfect" waveform but it does smooth out quite a bit with a 3hp motor idling with no load. I would be interested to see if anyone's unit is putting out a proper sine wave in reference to ground or if this is the nature of this design.
    Attached Thumbnails Attached Thumbnails pp-leg-3-ground.jpg  

  17. #95
    Join Date
    May 2019
    Country
    UNITED STATES
    State/Province
    Pennsylvania
    Posts
    195
    Post Thanks / Like
    Likes (Given)
    39
    Likes (Received)
    46

    Default

    As a side note I just wanted to document that there should be around 7 volts AC between Z5 source and Z5 gate connections on the board. The broken unit is outputting mV's

  18. #96
    Join Date
    Sep 2011
    Country
    UNITED STATES
    State/Province
    Virginia
    Posts
    25,989
    Post Thanks / Like
    Likes (Given)
    7561
    Likes (Received)
    8209

    Default

    Quote Originally Posted by Hardplates View Post
    This is very far from what I would call a "perfect" waveform but it does smooth out quite a bit with a 3hp motor idling with no load. I would be interested to see if anyone's unit is putting out a proper sine wave in reference to ground or if this is the nature of this design.
    Paul, yer a hard worker to have gotten back onto the unit after I bent your ear and yer Dad's. Good to have met you both.

    Yes, I think if you look at two things these switching artifacts are there from brand-new.

    - Jerry pointed out that whilst a P-P does some lumped-inductance fu cousin to a Scott-Tee to diddle two of the phases, it also has to do some VFD-style fu to convert 1-P energy and add a sine(isH) wave for the third phase.

    Elsewhere on PM there is more than one thread where "too much" of that noise got back up the "local" grid and messed-up microwave ovens and such on the same house feed.

    Now that I have but the white one, a 60 A 3-P transfer switch (it's twin for my gen set), can selectively put my 27 KVA EGS Hevi-Duty 230 Delta to 230 Wye Drive Isolation transformer on the clean sine wave of the RPC, OR the White-cased 10 HP P-P I bought brand-new.

    The extra filtering is but a bonus. The "real reason" is that I wanted a re-derived local Neutral.

    Upstream, I'm starting with an ignorant $80 Square-D add-on meant for protection from near-miss lightning strikes & such. "To Be Determined" if there is enough noise left to much bother with.

    If so, those other PM threads have a fair amount of discussions as to possible filtering alternatives.

    Short answer: For "Inverter Duty" motors, a P-P, output side, is probably harmless. For really OLD motors, "good enough". As-in still gentler than most VFD, even good ones.

    It's the UPSTREAM - the input side - and poorly-armoured (read "cheap PSU") household appliances where noise has been problematic, and even then, "only once in a while" - not all styles, ages, and individual members of the P-P tribe being equally noisy, input or "grid" side.

    3CW

  19. #97
    Join Date
    Aug 2014
    Location
    bainbridge island
    Posts
    1,129
    Post Thanks / Like
    Likes (Given)
    213
    Likes (Received)
    267

    Default

    Quote Originally Posted by Hardplates View Post
    As a side note I just wanted to document that there should be around 7 volts AC between Z5 source and Z5 gate connections on the board. The broken unit is outputting mV's
    I would have expected a bit more than that. typical to supply 15v to the gate to turn on and - something to turn off. i would expect at least 10 volts for turn on, -5 volts to keep it off. or it may just be +15 and -0 which your scope might read as 7vac.

    How did you measure .6 to .9 ESR for the output capacitors? That is on the order of 10 times as much as i would like to see a film capacitor of the order of 50uF 330vac rating. for example, a 60uf cap at 240vac 60hz is going to dissipate about 12 watts at .5 ohms esr, far higher than would be good. 1 ohm would be 25 watts . that's just the 60hz component. in this case you have the ripple current of the half bridge to filter as well.

    you can buy an LEM closed loop hall effect current transformer on ebay for about 30$ right now. you will need a +/-12v supply, a 10ohm resistor and two capacitors (to buffer the +/- supply at the ct), you can then run twisted pair from the 10ohm resistor back to your scope probe and you've got a 100KHZ bandwidth current meter. It won't be nearly as nice as a 500$ current probe given that you can't open the core and its only good for 100Khz bandwidth but you may find it handy.

    Can you read the part number on the gate driver for the old board or is it blown up?

    The parasitic spikes you are seeing on the output is just noise. however, if you can get a high pass filter to remove the 60hz component you could then look at the high frequency components of that spike and determine if its noise or if its a result of something wrong.

    You should be able to hook your Oscope up directly to the output of the igbts and see a square wave pulse width modulated smoothly from zero to about 80% duty cycle. if it suddenly goes to 100 or drops to zero around the locations of where that spike is on the output waveform then there is likely something wrong.

    you may need to buy a 100:1 scope probe if there is +/- 400 volts on the dc bus.

  20. #98
    Join Date
    May 2019
    Country
    UNITED STATES
    State/Province
    Pennsylvania
    Posts
    195
    Post Thanks / Like
    Likes (Given)
    39
    Likes (Received)
    46

    Default

    Quote Originally Posted by johansen View Post
    I would have expected a bit more than that. typical to supply 15v to the gate to turn on and - something to turn off. i would expect at least 10 volts for turn on, -5 volts to keep it off. or it may just be +15 and -0 which your scope might read as 7vac.

    I only threw a cheapo multimeter on it to measure the 7 volts which may or may not be true RMS at the switching frequency despite saying it is. I will put the scope on it in the future and get a more accurate measurement. I just wanted to post about it cause I use this thread as notes for myself sometimes.

    How did you measure .6 to .9 ESR for the output capacitors? That is on the order of 10 times as much as i would like to see a film capacitor of the order of 50uF 330vac rating. for example, a 60uf cap at 240vac 60hz is going to dissipate about 12 watts at .5 ohms esr, far higher than would be good. 1 ohm would be 25 watts . that's just the 60hz component. in this case you have the ripple current of the half bridge to filter as well.

    I tested them with a pretty inaccurate cheapo LCR meter whos readings I use as qualitive rather than quantitive. The old caps were giving the same reading as the new caps. I need to get a half good LCR meter at some point.

    you can buy an LEM closed loop hall effect current transformer on ebay for about 30$ right now. you will need a +/-12v supply, a 10ohm resistor and two capacitors (to buffer the +/- supply at the ct), you can then run twisted pair from the 10ohm resistor back to your scope probe and you've got a 100KHZ bandwidth current meter. It won't be nearly as nice as a 500$ current probe given that you can't open the core and its only good for 100Khz bandwidth but you may find it handy.

    What do I need a current probe for? I still have a lot to learn about electronics

    can you read the part number on the gate driver for the old board or is it blown up?

    No none are blown up, I have 2 broken boards and a running one. I don't know exactly what is the gate driver but I am assuming its a complementary pair of transistors. a MJD44H11T4G and a MJD45H11T4G. I believe these receive power from an UA79M08CKVURG3 -8 volt linear regulator. Don't take anything I say as gospel as it sounds like you know far more about electronics than I do. I'm still learning and relatively new to electronics


    the parasitic spikes you are seeing on the output is just noise, its pretty typical to see 10 times as much just from your oscope being near a vfd.
    See above red

  21. #99
    Join Date
    Sep 2011
    Country
    UNITED STATES
    State/Province
    Virginia
    Posts
    25,989
    Post Thanks / Like
    Likes (Given)
    7561
    Likes (Received)
    8209

    Default

    Quote Originally Posted by Hardplates View Post
    See above red
    Paul,

    What Johansen said.

    I use US-made - Pomona ISTR - & Fluke 100:1 and 1000:1 'scope probes. DC work, where a motor @ 300 VDC or less can kick 4 to 5 times that off the inductive spike.

    100:1 should do yah for the p-p.

    Motors, BTW, "integrate" or digest their food in the current realm. Ampere-turns as determine the strength of an electromagnet's field.

    Need of current measuring goods are because a 'scope is otherwise only a "Voltage sensing" animal.

    If nothing else, even when both have a pure Sine waveform - and they do not, always - Volts & Amps are near-as-dammit always out-of phase, "nominally" opposite, but again not-always exactly so.

    3CW

  22. #100
    Join Date
    May 2019
    Country
    UNITED STATES
    State/Province
    Pennsylvania
    Posts
    195
    Post Thanks / Like
    Likes (Given)
    39
    Likes (Received)
    46

    Default

    Quote Originally Posted by thermite View Post
    Paul,

    What Johansen said.

    I use US-made - Pomona ISTR - & Fluke 100:1 and 1000:1 'scope probes. DC work, where a motor @ 300 VDC or less can kick 4 to 5 times that off the inductive spike.

    100:1 should do yah for the p-p.

    Motors, BTW, "integrate" or digest their food in the current realm. Ampere-turns as determine the strength of an electromagnet's field.

    Need of current measuring goods are because a 'scope is otherwise only a "Voltage sensing" animal.

    If nothing else, even when both have a pure Sine waveform - and they do not, always - Volts & Amps are near-as-dammit always out-of phase, "nominally" opposite, but again not-always exactly so.

    3CW
    Copy, I will add a pair of 100x probes, a high voltage differential probe and a current clamp to my list of stuff to get. I seem to have expensive taste when it comes to hobbies or new interests.....but hey he who dies with the most tools wins.....right?


Tags for this Thread

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •