Horizon T81 Treadmill intermittent problem

Discussion in 'Technical Repair' started by Learn4ever, Jul 1, 2016.

  1. Learn4ever

    Thread Starter New Member

    Jul 1, 2016
    10
    0
    My treadmill runs normally but the belt suddenly stops at random intervals and then does a slow restart within a couple of seconds. A relay clicks when the belt stops. I have a hobbyist interest in electronics but am certainly no expert. I will appreciate any help available to diagnose this problem.
    The belt stops are becoming more frequent. They occur at the beginning and during an exercise session. Resisting the belt movement does not provoke the stop event. Lubrication did not help and there is no problem with bearings or drive belt.
    The JDYF02L RevG motor control board has no obvious problems. After reading several threads about this board, I have tried to probe the key components and all seem to be intact. Specifically the kickback diode is good and the MOSFET has high resistance between all terminals. I have not yet separated the board from the heatsink so cannot say if other components may be defective.
    Where should I start now? Although the three large caps have good physical appearance, is it possible for them to be the source of such an intermittent problem? Thanks in advance for all of you considering my problem!
     
  2. AlbertHall

    Well-Known Member

    Jun 4, 2014
    1,929
    381
    Intermittent connection - does hitting it provoke the problem?
    It may be an overactive protection mechanism of some sort, but in that case you're going to need a schematic.
     
  3. Learn4ever

    Thread Starter New Member

    Jul 1, 2016
    10
    0
    Yes-- intermittent issues are difficult. All the connections are solid. I could find nothing that would provoke the problem. However it may occur spontaneously at any time whether or not someone is on the belt or even if I resist the belt travel, so it is not a load mediated problem.
    A schematic could be of great help but one does not seem to be available. IamJatinah has shown great insight on other threads dealing with this board and I hope he will offer his thoughts about this issue.
     
  4. Learn4ever

    Thread Starter New Member

    Jul 1, 2016
    10
    0
    Here are some probing results that may help with the diagnosis. With the treadmill turned on but before the belt was started, AC input was 120V and input to the power bridge was 19VAC and the output was 18VDC, below the intended precharge value. With the belt running at minimum speed, bridge input was 120VAC, but bridge output fluctuated to be unreadable on my digital meter. I saw DC values from 50V to 100+V as the meter tried to sample the voltage. However, the output was obviously below the 160+ volts that should result from capacitor filter of rectified 120 VAC.
    Since the three caps are connected in parallel, they must be bad, even though there is no visible sign of overheating. I will post results of retest after replacing them. This would be a surprisingly simple cause for the problem when the treadmill works so well in other respects.
     
  5. IamJatinah

    Member

    Oct 22, 2014
    70
    14
    Learn4ever...great work! This is a bit perplexing and I am thinking on this...but here is what I can add...

    That +18vdc is correct for the pre-charge for the motor rail, as the start voltage for the 1mph at quickstart or Go will be about that voltage itself but with a pile of current behind it due to those 3 paralleled caps you mention, which are 560uf, 200vdc, chinese origin and yes get weak after 3-5yrs of work.

    The "Click" of that relay can be achieved from the lower boards control "PIC" chip, or the upper console board, both can remove power to the safety relay in the event the upper or lower sees any operational errors. So....... connections were verified and you sound well versed so I don't suspect things there, unless the connection in the Mast pole has an issue, as many times that console-to-MCB cable is a 2-part cable, with the same white 12-pin connector a servicer would see, but in that Mast there may be a mated black connector as well...I know...silly right? That's what I told Horizon when I worked for them, why create a failure point, designers eliminate such things.....so that's the only connections thought.

    The MCB, or lower board is one of my babies. If either the Kickback or the HexFet rip on this board, the other must also be replaced, it will at the very least have seen an extreme condition for it's ratings, so out it goes, and if either rips, likely takes out the other, and there will be no belt movement from that controller due to that safety relay, which will disengage if this board senses a short across the HexFet(IRFP260) itself.
    This board will also disallow that relay for bad on-board power supply(found at stand-up mounted 3A rectifier to circuit ground), shorted motor condition, low +V_Bulk(motor rail power....allowed from 125vdc to 177vdc approx, on those storage caps, the 3 in parallel), loss of upper "Enable Signal"(Active enable, pulled high thru entire operation of any belt movement), loss of upper speed command to MCB, or finally, any loss of either feedback sensor...being the front roller "slow" or encoder wheel "fast" pulse trains to that PIC chip.... ;o)

    Sooooo..... If you would like to contact me directly at jatinahATrocketmailDOTcom there is a device for testing these lower boards without the need for a console, this eliminates that possibility and allows a servicer to see, touch, feel that treadmill while in operation to check that lower board, rollers, motors, and feedback sources, but they are only available here, and we can talk further about that if you'd like. I can show you
    https://sites.google.com/site/jatinahelectronics/analyzer
    at this link what I mean by this....but basically I replicated the upper consoles functions across many platforms, like Matrix commercial treadmills, to Vision Fitness units, to Horizon Fitness units and then adapted some for Icon brands like Freemotion, Proform, Weslo, and others.

    There are no schematics that I know of, as there was one large set drawn out above my desk at that design job, which remained there, and no document control at this facility aside from don't provide any....good theory for a couple of reasons, but bad for the consumers.

    Good Luck! Please be careful working on motor controllers, and always bleed off large caps with a 100-ohm 5w axial resistor before touching.
    ;o)
     
    Learn4ever likes this.
  6. Learn4ever

    Thread Starter New Member

    Jul 1, 2016
    10
    0
    IamJatinah, thanks so much for the kind words and for the insight to the board. It has been so long since assembling the treadmill that the connector inside the mast was long forgotten, but it was snug and tight anyway. Good to check!
    So the question remains whether the caps are the key problem or whether the board senses another problem and turns off the power to the FET. Since the voltage reading on the bridge output was low and totally unsteady even when the relay was providing a solid 120 to the input, my guess is that the caps are the big problem. They should be coming next week.
    Many (!) years ago I had a tube type radio that popped often. All the tubes were good, voltages were right, but one of the electrolytics was bad. That is why I thought that a random problem might be caused by a bad cap. I will repost as soon as the caps are replaced. BTW, I have also ordered a kickback diode and a FET to be ready in case. ;) Your kind offer for a professional test of the board is appreciated and it may be needed if these attempts fail.
     
  7. IamJatinah

    Member

    Oct 22, 2014
    70
    14
    Sounds great...and I think you will find fresh caps will help nicely. There have been a few odd failures I have found over the past decade of working on these critters, that could cause problems but are tedious to find, usually freeze spray helps, sometimes I have found etching and trace issues, loose hardware soldered to unmasked traces, loose pins on SMT parts, intermittent 74HC14 chips(schottky inverter for wave shaping to/from the PIC), unstable on-board power supply, shorted bypass diodes on signal rails, pretty much seen most all circuits die on this board at one instance or another, and there were many hours put into this unit with both hardware and software ends, and after all the work and compared to other manufacturers, this little 1kw controller rocks right up there in the top 10% of those out there for this style of treadmill.

    I know that doesn't help fix the problem, but when this little fella is running, it's tight and can eat up to a 3hp motor in most cases.

    Remember there are THREE (3) separate "grounds" on this board, two of which are "HOT"(Off-line rectified, floating, includes motor rail and leads), a signal ground for the console and "human interface" portions(required to be at Frame Ground), and the on-board power supply Primary has it's own HOT ground as well, this helped to isolate noise from the main HOT grounded motor rail.

    Do not probe or scope across these grounding planes or you risk sending high voltages/currents back into your test equipment unless it too is isolated for testing, and those results are quite ugly(I am a former calibration tech and have seen the carbon fallout of this action, lol).

    Best of Luck.....we await updates when you have them ;o)
     
  8. Learn4ever

    Thread Starter New Member

    Jul 1, 2016
    10
    0
    New caps are now in place. I used 220 volt rated ones that were essentially the same size. The treadmill runs without any stop/ start issues, and the high voltage is steady at about 150. So it appear that my primary problem was indeed the three 560uf caps.
    The one remaining issue that I wonder about is seen only with a cold start. Starting the belt after power on occurs with a "one click" hesitation. In other words, I hear the relay click one (or more) extra time(s) before the belt starts moving. With a restart there is no hesitation. The precharge voltage is only 14 volts DC but with a restart the caps would hold a much higher voltage depending on how long the belt was stationary. Would the low precharge level cause this? What could cause the precharge level to be only 14 volts?
     
  9. IamJatinah

    Member

    Oct 22, 2014
    70
    14
    Hmmm..... There are 2 relays in there that will "click" at the beginning of the belt movement.....one smaller one is used to precharge the motor rail thru a small value/high wattage resistor or thermistor, then the second, larger "Safety" relay will engage shorting out the precharge relay/thermistor circuit out of the main motor rail as to not have that constant high current running thru a thermistor or resistor constantly.
    This "precharge" now charges those 3 input caps to a high voltage!!
    I know there is a delay from the second relay "click" to the time the motor starts to turn, but if you are hearing a double set of clicks, that might be related to a low Vcc to the console & onboard logic. I was thinking this Vcc was more like +15-17v but +14v should be ok for upper logic as there should be a > +2v barrier voltage between V_Console in and the first +12v regulator on the upper board, as there can be a steering diode in series which will drop .55vdc leaving an input of +14v - .55v = 13.45vdc which is pretty close to input requirements as most datasheets for the 7812 show a minimum input of +14.5vdc for regulation stability. It is not always good to assume a capacitor that is allowed to run in a circuit a few minutes, will gain or maintain more of a charge, as there are other factors in play with capacitors than simply a capacitance, as the internal ESR of that capacitor could be causing trouble, current drop thru cap, unstable holding capacity, erratic internal shorting at times. This is a puzzler......Output filter caps for the Vcc power are a pair of 820uf/25v which we replace with 1000u/25v.
     
  10. Learn4ever

    Thread Starter New Member

    Jul 1, 2016
    10
    0
    So let me understand the sequence of events on this board-- is the following correct?
    Power on: dc supplies come up for console and circuits on the motor board and also the precharge voltage on the motor rail through the small relay and the high wattage resistor. My board reads 14 volts across the three 560uf caps.
    Press start: Safety relay switches input of rectifier bridge to full 120 volts and turns off the small relay. Then the MOSFET is activated through the PWM circuit to start the belt. My meter shows an almost instantaneous jump to 150 volts but I have no scope adequate to show transitional events.

    The hesitation I see/hear with a cold start sounds like an on/off/on click of the safety relay although I cant test for sure. If there is a control mechanism that detects something wrong or if the motor does not get adequate voltage to respond at first would the PIC do this? Or is this normal and I never was aware of it? Thanks so much for your input! You have been most helpful!
     
  11. Learn4ever

    Thread Starter New Member

    Jul 1, 2016
    10
    0
    Well the intermittent bug strikes again. The belt stopped three times during my exercise this morning. It was a bit different since I did not hear the safety relay click in or out. The belt just stopped suddenly and then did a slow restart.
    The behavior now seems to be the controller responding exactly as designed for input signals that are out of range. But which input is the problem? The high voltage ripple was certainly a problem before replacing the caps. Now it must be something else. During my initial evaluation I checked the function of the sensor for the front belt roller, so it may be out of adjustment if there is any unique spec other than an on or off response as the magnet activates the switch. Can you suggest anything else that would be likely to provoke this response? Other than the sudden stops the treadmill works perfectly.
     
  12. Learn4ever

    Thread Starter New Member

    Jul 1, 2016
    10
    0
    The stop/start problem still occurs after replacing the filter caps and reworking the reed switch on the belt roller. A test for over an hour with constant belt speed of 3mph without anyone on the belt did not show a single stop. Regular use of the treadmill for less than a minute resulted in a stop however. So using the treadmill is actually dangerous and likely to produce a fall or a back injury.
    IamJatinah says that the control board will stop the motor if a fault is detected in the power supply, if the front roller reed switch is not activated, or if the console drops the enable signal. As far as I can determine these signals are now all working normally. Any other suggestions for why the control would stop the motor suddenly will be much appreciated!
     
  13. IamJatinah

    Member

    Oct 22, 2014
    70
    14
    Hi There, On the first post, mostly correct on the sequence of events except there are 2 relays that engage when the Start button is pressed....the first is the smaller one which allows rectified DC from the bridge to flow thru a small value resistor to limit the charge rate of the 3 capacitors you just replaced. This now energizes the 3 caps to about 150v from the bias voltage of +14v on that motor power rail....After about .7sec that larger "safety relay" engages which removes or "straps out" the first relay/resistor set now allowing full rectified V-Bulk to sit on those 3 caps and the motor rail is live and ready. This rail can then deliver up to +177vdc to that motor "switch", which in turn is rated at 200v(Vr_ds, breakdown)/ 35A (Id_cont)/ 46A (Id_cont_Cold).

    I have seen some strange behavior from treadmills and control boards when it relates to the motor integrity, as in Windings, Bearings, Shaft(wobble), isolation(from Frame Ground), brush-dust and recently I have seen 2 DC motors shorted so hard it bonded windings to the rotor shaft thru dielectric isolators but that would hold that controller off most likely.

    For the life of me, I have seen this very failure and I simply can't recall the issue, but thinking on this thru the past few days, I know I have run into this during on-site servicing a few years back before I stopped site runs.

    If you'd like an exchange unit let me know and I may have one I can exchange out for you, and we can get that miserable unit here, and dig in for all our sakes, or if you would like to proceed further, I have a couple thoughts...

    If continuing.....lets replace the 2caps, 820uf output caps in that onboard power supply, digikey part # 493-1305.
    While we are in that area of the PCB, very carefully replace R48 (220 ohm chip resistor) in the event this chip resistor is intermittently failing.

    Check or resolder that Bridge Rectifier DKB1 ( I Have seen this bridge blow open one of the four internal diodes, lowering it's capacity greatly and being a bit hard to see with it's mounting)......

    Also....very rarely I have seen carbon-deposits on the safety relay contacts cause issues at high current draws, passing current fine with no load, and then choking the motor rail when high current draws appear?....Usually this heats up that plastic case of the relay a bit as well.

    Lastly.....at times I have also seen treadmill circuit breakers get ugly at higher currents but this usually results in a premature bi-metal flip or "trip" of the breaker, although, anything is possible in this world ;o) Good luck, I will re-check-in soon!
     
  14. Learn4ever

    Thread Starter New Member

    Jul 1, 2016
    10
    0
    OK, thanks. That will take some time for an order. Meanwhile I would like to monitor the low voltage supplies. I see the 78M05 so that one is easy. What are the TP numbers for the 12 volt supply and ground? Or which components can be probed for the 12 volt supply?
    Since my last post I have pulled the safety relay to be able to check it and the bridge. The bridge is definitely ok based on multimeter ohm check. The relay also works well with 12 volts through the coil giving contact resistance of less than one ohm. However there is not much margin since a 9 volt battery does nothing to the relay. Since I had a new MOSFET I installed it just to be sure that was not the problem source. Retest after the above did continue to show the stop problem and one time during the prolonged stop there was a relay sound more like a growl instead of a clean click. The circuit was trying but the relay was not pulling in. So that could be either a bad relay or less than 12 volts from the supply.
    Does any of this remind you of the cause of the problem in your previous experience?
    To summarize I believe we have eliminated the high voltage supply bridge and caps but might have a problem with the relay since it seems to need a full 12 volts to pull in. The kickback diode and MOSFET are eliminated as the cause. The reed switch is ok and the motor and speed control seem to be fine. I assume the 820uf caps and R48 are part of the power supply so we are thinking in the same direction. R48 measures 216 ohms with my meter but I dont have a replacement right now. Do you have even a partial schematic of the 12 volt supply to help test it?
     
  15. Learn4ever

    Thread Starter New Member

    Jul 1, 2016
    10
    0
    New tests and same problem only worse. I spent some time tracing the board and l measured the supplies in the control section. The 5 volt supply out of the 78M05 was 4.9 and the unregulated supply feeding the relay was 15.7 volts from the cathode of D8(also at the anode of D1) to SGND. Replacing the 820uf caps made no difference in the readings or the stopping issue although the first attempt after the work resulted in several seconds of relay chatter before the belt started moving.

    The relay is controlled by Q2 adjacent to D1 which is across the relay coil. I observed the base drive to Q2 fall to zero which opened the relay and caused the belt to stop. Since this sequence is likely originating from the control software, the question is what signal is faulty to cause the response? The high voltage circuit seems to be ok and the control section power supplies look ok. The reed switch checks out and the motor speed sensor appears to be working ok. Maybe the motor control board is ok and the problem is somewhere else? I'm not sure where to go now. Any ideas?
     
Loading...