I recently finished repairing a Dynapath Delta 40M control on a Chevalier Falcon knee-mill and would like to share my experience so others can avoid the mistakes I made.

The control showed signs of impending failure years ago. I'm told that the operator wasn't able to load a part program from the control's internal memory. Likewise, new programs wouldn't save. Our shop has 4 VMCs and an HMC, so the little knee-mill was abandoned from use and sat for 2-3 years. But as my company's first CNC, I think this machine has a special place in my boss's heart. So I was asked to get it working again.
When I powered it on, the first thing I noticed was that the system clock was not set and the AT keyboard must be connected to get past the bios post. We didn't have one nor did we have an AT->PS2 adapter, so I couldn't start the control. The PC mainboard in this control has a non-replaceable 3-cell Ni-CAD battery which is soldered to the board. So being that the machine sat for years, this error might not mean the battery is shot. Leaving the power on for an hour or so could fix this problem by allowing that battery to charge. Not the case for me: Upon opening up the control, I saw this battery was covered in crusty white crud which is a good sign that its shot. I removed the mainboard and de-soldered the old battery. I found a replacement of the same style at a local electronics store with the only difference being that the new one was Ni-MH. Ni-MH and Ni-CAD batteries are trickle charged the same way (quick charging however requires different peak-detection methods), so it was a compatible replacement. I soldered the new battery on and it solved that problem. I still had to get an AT->PS2 adapter to set the system clock, though. For those not familiar with the AT keyboard interface, it is a standard PC keyboard except the connector is a full-size 5-pin DIN connector whereas PS2 uses a 6-pin mini-DIN.
With that out of the way, I was able to start up the machine. Next problem: After booting up into the control's OS, I received an error message on the first screen. It said "Touch soft-key invalid or not present" and none of the panel keys seemed to work. The 7 software buttons below the monitor did however work. I thought that the "key" referred to in the error was the panel keyboard and wasted several hours swapping the key-panel out with one from a working Delta-20 on another machine. I called Dynapath and found out that the "key" referred to in the error is a hardware key-chip located on the servo transducer board. It contains the control serial number and locks the control software (which also has the serial number) of that particular machine to its hardware. So the control software from one machine cannot be used on a different, yet identical, machine unless this key chip is swapped as well. The key chip looks like a thick button-cell battery and Dynapath told me they've had countless people mistake it as one and fry their ST board by replacing it with a battery of similar size!
After telling a Dynapath tech my control's symptoms, I was told I needed a new key chip. It cost $120 and I had to tell them the control's serial number so it could be properly programmed. Here's a little more detail about the symptoms I was having:
1) Error message on first screen: "Touch soft-key invalid or not present"
2) CNC executive ID displayed all ???????. (This means the control can't read the serial number from the key chip). PLC ID number was still there.
3) All front-panel keys did nothing except bring the contol out of standy.

I played with it while I was waiting for the new key chip to arrive and found more symptoms. With the control cover removed I powered-on the machine and noticed that sometimes the watchdog timer LED would come on immediately and sometimes a few seconds after releasing the E-stop. This LED is on the I/O board near the bottom. I think it was labeled something like "com/timeout". Anyway, if this LED is lit, it means that the PLC has lost communication with the rest of the system and the E-stop is set (regardless of whether the E-stop switch is in or out). Every 5 seconds the I/O board sends a short message (ping) to the PLC (which is located in the control pendant behind the key-panel) to see if it is still there. If it doesn't receive an echo ping back from the PLC within a certain amount of time, communication is stopped, E-stop is set, and this error LED lights. This check circuit is commonly known in the realm of electronics as a watchdog timer. When the LED wasn't lit, releasing the E-stop would cause the servos to come on and hold. But this never lasted for more than 2-3 seconds before that watchdog timer LED would light and disable everything.

If your watchdog timer LED comes on somewhat randomly, it is a hint! If it always comes on at the same time when you do something, a different issue is implicated. But when it lights somewhat randomly, it means you either have random EMF interference being picked up in the RS-232 cable between the I/O board and the PLC, or you have hardware which hasn't completely failed but is teetering on the verge of failure. A bad connection of the RS-232 cable will more likely present itself consistently, but its an easy thing to rule out by cleaning the connectors with contact cleaner.

When I received the key chip, it only took a couple minutes to install it and discover that it did not solve my problem. At this point I figured I didn't have the resources to test each of the control's boards, so I sent the control to Dynapath. They didn't require the PLC, so it was easy: Just disconnect all the wires, remove the four screws holding the control box to the main enclosure and ship in a good container or box.

It turned out that the PC mainboard was at fault. The ISA bus which all the Dynapath boards use to communicate with each other, was not working properly due to aging components. That explains why the machine operator wasn't able to load a part program: The control stores all the parameters, setup info and part programs on an ISA card called the memory board. Data wasn't getting through the ISA bus from this board at that point. By the time I started messing with it, it further failed to the point where the I/O board and the servo transducer board (containing the hardware key chip) couldn't communicate, thus the "??????" for CNC executive ID. BTW, Dynapath told me that when this is the case, all functionality is disabled by design. So if you have those question marks, don't expect to be able to get anywhere by messing with it, because at that point, the control has locked you out.

A new replacement mainboard from Dynapath cost $895. I believe that if you are strapped for cash, this board could be replaced with most any PC mainboard circa 1994, but it does have some special features that aren't necessarily required. One such feature is the CMOS battery which is rechargeable whereas typically these are non-rechargeable CR2032 lithium button cell batteries. Don't expect a lithium cell to last the 16 years that the rechargeable one did. 5 years is about average, but they're easy to replace so it really isn't a big deal. The other special feature is the expansion bus slot layout. Any substitute mainboard is going to need at least 4 ISA slots. The mainboard used in this control has what looks like PCI slots butted up to the end of a few of the ISA slots. None of the control cards in my machine used these, and I believe their presence means this is a server board. The only things I've ever seen to use this ISA and PCI-like slot style are processor boards in servers where you can add processors to your system by simply inserting another processor board into one of these slots. Don't know, its kinda from before my time. For sure though, unless you find a PCI to ISA expansion card, the newest mainboard that will have 4 legacy ISA slots will be from a Pentium III or equivalent system. I'm pretty sure ISA was completely phased out when the P4 processor was introduced. Regardless, the original processor in this control is a whopping 33Mhz, so a PIII at up to 1.2Ghz is wayyyyy more than enough processing power for running NC programs. You'll never even see any benefit from a more powerful processor than the original except perhaps in decreased boot time.

Back to my machine. After Dynapath diagnosed and replaced the PC mainboard, my machine now worked.... until I tried to reference home. The following is a completely different issue and has nothing to do with the control, but everything to do with diagnosing servo problems.

When the y-axis was commanded to find the limit switch during the reference-home process, it would jump back and forth rapidly and run-away until I hit E-stop or until the control realized it had lost control of the servo. I took the servo out and when I removed the orange end-cap, way-lube poured out! Not good. I cleaned the living piss out of the encoder and the rest of the servo as well as I could without further disassembly, but the problem persisted. I was about to send the $400+ encoder to a professional repair place, but decided to further disassemble the servo first. What I found was all four of the small brushes in the back were broken. Probably from sitting saturated in oil. This rotor and brush setup in the back of the servo is the tachometer generator. It generates a voltage when the servo is running which the control uses to determine if the servo is running at the correct speed. If the axis is loaded, the servo RPM will drop unless the power is increased to keep it in sync. The tach-gen voltage is used by the control to sense this and keep the RPMs exactly where they should be under varying loads.
And it was the broken brushes on the tach-gen that were causing the axis to jump around wildly. Replacement brushes were expensive as far as brushes go: $11.20 ea from a local electric motor repair shop. But it was a lot cheaper than the cost I would have incurred if I had sent the encoder for unnecessary repair (was gonna be a $120 minimum charge).

So now the machine and control work perfectly and the machine is back in use making parts. One nice thing about this machine is that with 5 machining centers at our shop, none of the operators want to use it. So its pretty much always available for tool and setup making purposes (and personal jobs).

I hope anybody seeking help with this control is able to make use of my experience and I welcome any questions you might have. If you have any questions I might be able to help with, email me at [email protected]

PS, I personally own a Supermax YCM-40 with a Centroid M-50 control and have experience upgrading the internal PC and performing other repairs, so if you've got questions regarding this machine I'm more than happy to help.