[rlarios]

This is something that I’ve had happening blowing 2 – 100A input fuses out of 3 at least once every summer during the last 4 years and/or occasionally turning LED 2 on early in the morning when machine is first turned on the rest of the year. This LED 2 alarm doesn’t happen once the machine has warmed up to operating temperature.

It happens that I’ve read postings in this forum about this issue saying that it’s caused by some hybrid ICs in the SCR firing section which have to be warmed up prior to running the machine. By the way, the LED2 alarm happens while the machine is idle waiting until it’s at running temperature. According to a few pages I once got from Fanuc World, if LED 2turns on, one has to jumper CH16 to 0V. If LED 2 goes out, the problem is a field loss, if it stays on, it’s an overcurrent problem, which is my case. See attached pdf file.

Anyway, the machine could run most of the time until about 4 weeks ago it stopped and the Spindle board turned on LED 4. This is something I had never seen before. Apparently, the Spindle motor temperature sensor was activated but it was not. The LED 4 turned on as soon as I flipped the circuit breaker on, so something else was going on there. I removed the Spindle PCB from the machine and as I had no schematics whatsoever, I decided to see what caused LED 4 to come on in the first place. Soon I realized that the cathode of LED 4 was wired to M5 pin 1 which was shorted to ground. I kept looking and found that M7, M6 and M1 were also shorted to ground. MX designation for this board is for TTL ICs, a bunch of NAND and NOR gates.

A damage of this magnitude was not something that just occurred randomly as this board had survived more than 30 years without a problem. After several hours of reverse engineering this board, I found that the“fuse-blown” section of the Alarm fuses was connected to M7 IC pin 4. Checking the alarm fuse corresponding to F1, it was blown inside and it looked as if the 1.3A fuse arced inside the cartridge which also arced with the DC section that goes to M7. This arcing in turn took with it the rest of the ICs mentioned above. After replacing these ICs and clearing the shorts, I decided to put the Spindle board back into the machine and turned it on after having replaced F1 as well. When turning the machine on, LED4 stayed off, which was a good sign.

Then I had the operator test the spindle (after the mandatory one hour warm up time) by running it forward at 500 rpm, but the spindle did nothing. No lights on, no nothing, then I decided to continue to reverse engineer the board until I found what caused it to not run when commanded to do so. After some more hours, I followed the CN1 connector signals until I found where they went. Soon I found that control signals go into some square black ICs (A-RV03) that seem to level shift the signals from 0 – 24V down to 0– 5V for TTL logic. As the operator tried to run the spindle forward, I decided to re-install the PCB back to see with an oscilloscope how the signal behaved. Soon I noticed that the SFR signal (Spindle FoRward) went from 24V down to 0V and the output went from 4V to 4.5V which didn’t cause the spindle to run. I asked the operator to run the spindle in reverse and this time the spindle ran. This proved that as the SFR signal was active when the arcing occurred, the (HY4-IC) A-RV03 line corresponding to this signal was blown. In order to test for the rest of the circuitry related to SFR, I removed the jumpers that deal with spindle rotation on CN9 and rewired the pins so as to make the spindle run forward when commanded to run in reverse. Sure enough, it ran forward. This meant that only the HY4 IC was bad.

As I had found a couple of years ago that user jolulank mentioned something similar (the blown fuses thing), I asked him where he got the A-OS04s and A-OS03s ICs he said had been replaced in his machine and he said they couldn’t easily be found. The only option was to buy a junked board and pull out ICs as they were needed. So, we did, we bought a junked PCB for the parts. This time, however, the bad part was an A-RV03 which was recovered from this spare board and installed back into the original PCB. Unit was tested and Spindle ran in both directions. Now we could go back to original problem.

LED 2 and/or blown Fuses:

Jerry Shorter posted the following regarding fuses blowing:

https://www.cnczone.com/forums/fanuc/95354-input-fuses-blowing-6t-spindle-drive-2.html

HRH replied: “You possibly need a technical person who knows the drive and can check the firing and current signals with an oscilloscope.”

jdeere replied: “How is the accl/deccl is it rough, to aggressive?
Does it run smooth at commanded speed?
Possibly if either hybrid chip on the pcb replaced or the dither adjusted.”

In another posting, Jolulank asked about the same in his Fanuc 5T:

https://www.cnczone.com/forums/archive/index.php/t-113496.html

AI_The_Man replied: “Quite possible the firing board is causing the problem.”

Tmbruno28 replied: “firing problems can be check as follows:

Check in CH17(a, b, c) with an oscilloscope. If wrong exchange AOS04.
Check inCH18 (a, b, c) if wrong, exchange AOS03.
If problem occurs not at each time, first exchange AOS03.”

Armed with all this and Fanuc DC Spindle Maintenance Manual B-51649E/05, I checked the adjustments and found at CH17 a, b, c (synchronous pulse balance) that all three signals were about right per the manual while spindle was stopped. At 60Hz, the pulse widths measured about 1msec. Manual recommended between 0.8msec to 1msec.

However, measuring at CH18 a, b, c (dither shift circuit balance), all three showed 2.7msec pulse width while the manual recommends 1.4msec. See attached photo.

It was at this time when I called a CNC man based in North or South Carolina who told me that he could fix my Spindle board if I wanted, but to never touch the adjustments otherwise the repair would be more time consuming and expensive.

This made me hesitate, however, reading all the postings about this Spindle control made me wonder whether adjusting the settings could cure my problem. The CNC man mentioned, however, that 9 out of 10 Spindle boards with same symptoms have problems with Power Supplies. The only issues, so far, that I’ve seen are that +15V supply varies from +14.90V when cold down to +14.86V when warm, and the +24V supply varies between +23.65V to +24.75V either cold or warm, and shows more than 1V ripple voltage. The other supplies don’t show noticeable ripple voltage.

Another thing, the posting made by Tmbruno28 made me think, what did he mean by “if wrong” replace AOS04 or AOS03? In the case of “dither shift circuit balance” should I change all three AOS03s? Are adjustments correct at 2.7msec or should I adjust them down to 1.4msec if possible? All three settings a, b, and c have practically same pulse width.

If anybody has experience with these issues, please let me know whether I should either adjust or replace ICs. Also, if anybody can share or sell the schematics for this Spindle board I would really appreciate it.

Rafael

[generaldisarray]

have you considered a ac motor/drive swap?
I did one on a lathe with a Fuji dc drive. cheap and easy

[rlarios]

Yes, we have. However, motor is still good and spindle runs very well as long as we let the spindle drive run when warm. Fuse blowing occurs only when we let machine run cold which happens only when operator forgets to let it warm up.

I wish I could get some feedback from some of the experienced users which have mentioned words like "dither", or "if signal wrong do this. " etc. as they seem to have gone through same process. If I don't hear from anybody else I will have no choice but to go ahead and start by adjusting dither shift circuit balance. If for any reason I can't adjust dither, I will replace ICs until I get it right.

HRH -he seems to be in South Africa- mentioned in one of the threads I quoted in my first message:

"A local company has made crude but very efficient replacements for the hybrid IC's on the control PCB. I have solved several intermittent LED2 (Over-current alarms) and blowing input fuses with the re-engineered units.

In some cases the hybrids need to be resoldered which in turn "resolders" the internal connections of the pins to the internal pcb fo each hybrid. The hybrids have a serious design flaw when it involves expansion and contraction due to temperature changes.

The success rate is very high as long as the drive is also cleaned and checked thoroughly."

I tried to reach him a couple or so years ago but I was unsuccessful.

In actuality, his comment is the reason why I am insisting in getting this fixed. I just wanted either a confirmation, and/or recommendation from the group. The schematics of this Spindle board and/or these ICs would be very helpful in putting together my own "crude" replacements, though.

Nevertheless, I appreciate your input generaldisarray. Your suggestion will be the next step to follow if nothing works.

Thanks again.

Rafael

[rlarios]

Duplicate

[rlarios]

*** PROBLEM SOLVED ***

Given the fact that apparently nobody works with these Fanuc DC Spindle boards anymore, I had no choice but to plunge in into the problem.

I started by checking the adjustment which apparently was wrong in all three phases (CH18a, CH18b, and CH18c). It turned out that thanks to the fact that I had downloaded from somewhere a Fanuc 6T maintenance manual, I noticed that in order to check this adjustment, connector CN2 had to be removed and CH13 (+15V) had to be shorted to TSA (tacho-generator signal test point). Sure enough, doing this reduced pulse width from 2.77msec down to 1.4msec just as the manual said it should. Being this the case, no adjustment was necessary. I later discovered that this note was also in the Fanuc DC Spindle Servo Unit MaintenanceManual which I obviously had overlooked.

Next, I continued googling for DC Spindle schematics and Ifound a site in Lithuania that sells CNC schematics:

https://cncon.lt/wp/?product=fanuc-a06b-6035-cxxx-wiring-diagrams

At the time I got there, the site was not secure. Now it seems it is. I went through the entire document and used the last 30 pages out of a total of 100. The document was not easy to read, as it is rotated 90 degrees. So, I placed my laptop on its right side so I could read the schematics. Sure enough, schematics were older than my Spindle board so I started to compare both and they were close enough to work with.

Not knowing what to do, I first went to find out whether LED2 got turned on by either an over current or field loss alarms. I used my oscilloscope and configured it so it could acquire a single event such as these alarms. It happens that once the alarms occur, they set a flip-flop which turns on its corresponding LED which cannot be turned off until Alarms are reset, except for LED4. The Fanuc 6T maintenance manual shows a map of all CH test points as well as the Auxiliary test points which the other manual doesn’t show. Everything made sense once I got these schematics.

First thing first, I had to wait until the machine was cold. I had also to wait for a reduced work schedule on this machine. Once I got there, I set up my scope placing the test probe on the auxiliary test point “T”which according to the manual, corresponds to the “Over Current” signal. Then,I started the machine and let it sit idle, this means, Spindle stopped and waited until the error occurred. Sure enough, in less than 5 minutes the machine tripped and I went to see what the scoped grabbed. There it was, the overcurrent signal that sets LED2’s flip-flop, see Photo 1. This proved LED2 was set by an Over Current and not a Field Loss event.

I turned the machine off and let it cool down for about half an hour before I could do another test. Checking on the schematics I realized that this OC signal came from a comparator, and a series of Op Amp stages all the way back to the first stage which was a differential amplifier whose inputs were the Armature Current Sensor. The output of this Op Amp can be monitored on CH24. Manual said that this signal shows about 5mV (or 20mV, it depends on what manual you see) for every 1 Amp of Armature current. After three more Over Current trips, I couldn’t grab this overcurrent coming from the diff amp, and puzzled, I let the machine operator continue his work.

Next time I decided to use this OC pulse as a trigger to acquire CH24 signal. Well, the reason I couldn’t grab the signal from the diff amp the day before, was because it was inverted and I had assumed it was not. See Photo 2. At this point, it was weird as to why there was current flowing through the armature when the Spindle was stopped. Thinking that maybe something else was going on, I placed a DC current probe on one of the wires that goes to the armature and waited until machine tripped. There it was! There was current flowing through the Armature. Something must be turning on the SCRs which causes this current flow in the first place. See Photo 3.

Then I placed an AC current probe on Phases A (U), then B(V), and then C (W), one at a time and found that at the moment of overcurrent, Phases A and B had abnormal current flow. Phase C didn’t show any abnormal current flow during OC trip. See Photos 4, 5 and 6. With this information at hand, I went to the schematics where SCRs are shown, and noticed that in order for this to happen, SCR1-SCR4 or SCR3-SCR2 had to be on in the Forward direction. SCR8-SCR9 or SCR7-SCR10 had to be on in the Reverse Direction. These SCRs combinations could be turned on with signals UFP, UFN, URF and URN. All these signals came from an A-OS03 IC. This A-OS03 gets controlled by Phase A (U) synchronization voltage going first to A-OS04 then to A-OS03, SIGNF (Forward Direction signal), SIGNR (Reverse Direction Signal) and ER (Phase Angle Command Voltage). I tested these signals at Over Current time and nothing happened. See Photos 7, 8 and 9.

Then, I decided to confirm whether this A-OS03 IC was indeed bad, so I soldered wires on solder side of Spindle board going to the signals UFP, UFN, URP and URN to test them with the scope. Here I have to make clear that after the control tripped, I had to turn the machine off and wait for at least half an hour to let it cool down for next test. Sometimes, the machine didn’t trip and I had to turn it off again and wait for at least one full hour.

Well, to make a long story short, UFP, UFN and URP didn’t fail. At this point I was nervous as I had only one signal left, but thank goodness, URN failed and I was able to grab it. See Photos 10, 11, 12 and 13. This A-OS03 was removed and replaced by another A-OS03 from a junk Spindle board that was specifically bought for this purpose. See Photo 14. Machine was tested and the problem seems to be gone as it hasn’t tripped since it was repaired 4 days ago.

I am posting here what I did so others with similar problems can fix them. This Fanuc DC Spindle board is an engineering marvel. It operates with Operational Amplifiers, Comparators, TTL circuits and a few 555 timers.They still work in spite of being around for more than 35 years!

I went through all this because I didn’t know where to start. Fanuc experts may chuckle, though. Now I know what to do next time. This proves also that a faulty A-OS03 can turn on SCRs at the wrong time which may cause fuses to blow. I can say now that this problem is SOLVED!

Rlarios

[generaldisarray]

nice work :cheers:

[Nicotronic]

Hello rlarios. Very good job!

I'm a french technician.
I works on the same problem. Overcurrent alarms occurs some times. I have scopped ch18 ch17 on each a,b,c. I can see that the signal ch18b and ch17b lost theirs signals simultaneously sometimes.... I,'will try to exchanged aos3 and aos04 soon. But I would know where are you find the signal ufn urn ufp and urn? In a schematic?
Thanks a lot for you reply and apologize my poor english
Best regards

[rlarios]

Hello rlarios. Very good job!

I'm a french technician.
I works on the same problem. Overcurrent alarms occurs some times. I have scopped ch18 ch17 on each a,b,c. I can see that the signal ch18b and ch17b lost theirs signals simultaneously sometimes.... I,'will try to exchanged aos3 and aos04 soon. But I would know where are you find the signal ufn urn ufp and urn? In a schematic?
Thanks a lot for you reply and apologize my poor english
Best regards

Hi there! Good to know that this thread was read by somebody!

Yes, those signals are shown in the schematic. I haven't visited that site since then, but if you can't find the schematic let me know and I will post the PDF file I got where you can see everything you need.

Rlarios

[phucrobusl]

superb work ???????????? post the pdf as soon as possible.

[Machineguy]

Hi, you say you replaced the motor and drive to a fuji AC unit from a DC unit. Spindle drive only or did you replace the X axis And Z motors and drives as well. What was the cost of the parts for this, you say cheap?

[rlarios]

Hi, you say you replaced the motor and drive to a fuji AC unit from a DC unit. Spindle drive only or did you replace the X axis And Z motors and drives as well. What was the cost of the parts for this, you say cheap?

Hi there! I just re-read the entire thread and didn't see anything related to what you're mentioning above.

However, let me tell you that it's been 2 full years since this thread was started and machine is still going strong.

I guess I will bring a cake to celebrate the second anniversary of this repair to the shop.

I just noticed that someone requested a copy of the dc spindle drive schematics. I will post it tomorrow, though.

Have a great day!

ratios.

[rlarios]

Hi there! I just re-read the entire thread and didn't see anything related to what you're mentioning above.

I stand corrected. I noticed that generaldisarray suggested to do a motor/drive swap after the first posting. You should contact him for more information.

rlarios