Protecting electronics on the XS650 from voltage spikes.

I know you believe that you fried a Tytronic ignition, and that is what spurred this whole line of thought. Have you heard of other electronic ignitions being fried by wayward voltage spikes? It makes you wonder why ignition manufacturers don’t provide some sort of “ insulation “ when they make them.
 
I know you believe that you fried a Tytronic ignition, and that is what spurred this whole line of thought. Have you heard of other electronic ignitions being fried by wayward voltage spikes? It makes you wonder why ignition manufacturers don’t provide some sort of “ insulation “ when they make them.

Pamcos were notoriously sensitive to high primary voltage, PMA, cap only, cheap chinese regs that failed in full charge mode, what could go wrong? Pete said yeah I can add protection but then they would be more expensive.... I think this contributed to his getting out of building.
TCI's tend to DOA when any arc welding is done....
It's likely other brands include some sort of protection then again I hear they moan about physical location, worrying about EMF from the secondary. Pete chewed me out for shoving an early E-advance under the gas tank next to the coil.
The one on Madness lives behind the backbone under the seat.
 
Last edited:
The TCI box has multiple instances of using a combination of caps and diodes on the power line to dump spikes to ground.

TCI - Schematic.jpg


I remember back in the day... when IC's were fairly new tech in aviation, we'd get several change orders (TCTO) a year to open a box and add a cap or diode somewhere near a chip that was failure prone. To give you an idea of what it was like back in the early 70's... F111D's had HUD's (heads up display). When it first came on line, the HUD signal conditioner had a MTBF (mean time between failure) of about 2.5hrs iirc. The average sortie lasted 3.2hrs.... so they were failing on just about every flight. We did lots of mods... newer chips as they evolved and always adding caps and diodes. When I left the program 2 yrs later, we were over 100hrs MTBF. That's a big improvement in anybody's book... but compared to today's boxes that go thousands of hrs, they weren't very reliable.
 
The discussions about decoupling caps show both a large cap (for lo-freq surges) and a small cap (to absorb hi-freq noise). Those are for hi-tech computer gadgets, running signals and clocks in the hundreds of megahertz to several gigahertz. For the lowly XS ignitions, we can just deal with much lower frequencies. I found that the Kettering ringing on the ignition coils to be around 10khz.

As xjwmx mentioned, it's best to look at the power lines with a scope, see the sharpness and magnitude of the line noise. Then add suppressor caps as necessary.

Flying blind, I could suggest adding 47uF 25v electrolytics to the ignition coil's power side, and at least 1uF 25v caps at the power inputs to other devices. The only way you'll know if this works, is if stuff doesn't release smoke...
 
The TCI box has multiple instances of using a combination of caps and diodes on the power line to dump spikes to ground.

View attachment 176902

I remember back in the day... when IC's were fairly new tech in aviation, we'd get several change orders (TCTO) a year to open a box and add a cap or diode somewhere near a chip that was failure prone. To give you an idea of what it was like back in the early 70's... F111D's had HUD's (heads up display). When it first came on line, the HUD signal conditioner had a MTBF (mean time between failure) of about 2.5hrs iirc. The average sortie lasted 3.2hrs.... so they were failing on just about every flight. We did lots of mods... newer chips as they evolved and always adding caps and diodes. When I left the program 2 yrs later, we were over 100hrs MTBF. That's a big improvement in anybody's book... but compared to today's boxes that go thousands of hrs, they weren't very reliable.

Yes Sir I understand that you know these things Jim
Firstly I Don't know electronics .
I have looked at the Schematic and it appears to be the electronics of some box .fairly new .. since it is dated this year
I have never seen a schematic like this for the internals. And those are rarely open so one can repair or understand .Often parts are protected by some
plastic Epoxy ??? ( Regulators )
I am not after infringement on someone elses intellectual property.. This can be something that is a part of a Business. Or so
And not allowed to use
If not I find it interesting and Would appreciate if more information can be given .
There appears to be a Controller 20 pins with markings that I cannot find on the net .. Programmable Curve ??
Is this something tested and in Use.
Thanks Jan
 
On that circuit diagram the 20 pin chip is a quad comparator. This came in a couple of versions. On my previous 81 XS this component was a custom item made by Hitachi and contains additional circuitry, but unfortunately it seems to be no longer available. In some other XS models I have seen, the TCI quad comparator chip is non-customised and easily replaced with a modern version. The good news is these comparator chips seem to last forever.
 
Last edited:
On that circuit diagram the 20 pin chip is a quad comparator. This came in a couple of versions. On my previous XS81 TCI this component was a custom item made by Hitachi and contains additional circuitry, but unfortunately seems to be no longer available. In some other XS models the I have seen the TCI with a quad comparator chip non-customised and easily replaced with a modern version. The good news is these comparator chips seem to last forever.

Thank you Sir
Now I can read about it
My first thought what would happen if replacing that comparator with a modern Microcontroller one that is Programmable.
Into that same schematic with some lever adjusting extra electronics
For different curves depending on insignal from rpm.
The comparator has a specification somewhere and the modern controller also .
Am I thinking wrong here ?
 
I would suggest making a simulator to trigger the stock TCI on the bench at various rpms and monitor the output using an Oscilloscope to see just how this circuit works. Armed with this knowledge you could easily make a new circuit board with modern high quality components...


...or jump to the thread on the GN250 ignition system and save some time and trouble.
 
As a matter of interest, took a peek at my '74 TX650A.

Red is AC probe from one of the 3 white stator wires. Blue is AC probe from the power bus at fuse box (only shows the variance around the 13.5v DC it was idling at).

Red shows stator events and blue shows the ignition events (including dwell time when voltage starts sinking):

upload_2020-10-17_14-14-5.png


Can see time markers are set at ignition intervals and show 58.61ms. 1000/58.61*60 = 1024 RPM. Powerbus voltage variance is 0.21v without any noticeable spikes. That variance from ignition events is pretty normal from my experience fooling around with automotive stuff, but normally I see more erratic spikes. Based on all the talk about how noisy these bikes are, I was a bit surprised it was this clean.

Anything can make it ugly, mostly bad grounds. My wiring is all new, new brushes, VR1010 regulator, new solid state rectifier, original stock rotor/stator, Pamco ignition.

IMO, it's not really worthwhile adding components to try to clean up the power. KISS. Just make sure your wiring is proper. Good clean solid grounds, hub/spoke power distribution, and good clean solid grounds. <- intentional redundancy lol. If you add an electrical component to the bike, most well made stuff will already have the circuitry to clean up what it needs. It's just not hard or expensive for a manufactured product. But when fooling around with homemade circuitry, a buck regulator is a great way to clean up power, and a simple RC filter for signal conditioning, but don't guess at it because the capacitors create delays which may or may not impact the timing of the signal you are conditioning.

I'm not a pro, just fooled around a lot with Megasquirt projects about 15 years ago. A true rabbit hole. Few arduino projects too. Hobby scope is a great thing to have. Also have a DC current clamp to plug into it which has been useful to see on a scope.
 
I would suggest making a simulator to trigger the stock TCI on the bench at various rpms and monitor the output using an Oscilloscope to see just how this circuit works. Armed with this knowledge you could easily make a new circuit board with modern high quality components...


...or jump to the thread on the GN250 ignition system and save some time and trouble.
I made a test jig using an Arduino UNO. It is a bit crude right now but it is useful for testing TCI boards. I intended to post more info after I had refined the project more. I have also started another project to build a new TCI module using a Arduino nano to measure the rpm, and trig the ignition coil using an algorithm. The problem that I am running into ( Besides the fact that I have to many competing projects on the go) is that the Arduino is only capable of doing one thing at a time, especially when it comes to measuring time. So far I can measure the time between the second pickup pulse to the first, apply an algorithm and then fire a signal that would coincide with the appropriate ignition firing. But I also need to do the same steps to charge the ignition coil( dwell). Can’t do both at the same time, at least not with my limited programming skills. Will post more when I can.
 
Last edited:
The TCI box has multiple instances of using a combination of caps and diodes on the power line to dump spikes to ground.

View attachment 176902
Jim,
Yah, I worked on a lot of old RADAR equipment from the 60's. It was the early days of solid state equipment. Lots of germanium components that would fail if you looked at them crossly. That stuff we had was supposed to be mil-spec but that didn't seem to mean much. Far more delicate components, than they make today.

Just wanted to add a couple of notes on the diagram Jim posted. The pickup pulse signals coming into the board are negative-going....thus the orientation of the input diodes D1, D2, D3, D4. The input transistors are normally on, except when the pickup pulses come in, then they are switched off. D5 and D6 also act to shunt any negative spikes to ground, but basically they just clip everything to about -0.6v. I believe the resistor/cap combination on the input also acts as a filter to prevent spurious noise from coming in. I suspect there may be some engineering-type-thinking going on here, in that it may be easier to filter through fairly large negative pulses amongst a lot of rotor induced noise ( a lot of which may be positive). I have never been able to effectively measure the noise on these lines. A lot of the circuitry on the board runs on the regulated 8 volts provided by ZD1 and C1. There can be all kinds of crap on the 12 volt bus, but if this regulator on the board is working, then the board timing circuitry should be OK. In theory the board should work with a battery voltage all the way down to 9 or 10 volts, but by then you won't be getting any umph out your ignition coil. I believe the zener diode ZD2, C2 and R3 are part of a startup delay circuit type circuit. Basically when the board starts up you want the circuitry to start up in a known condition and maybe not randomly fire the ignition coil. Without knowing more about the IC, it is hard to say, but there is something familiar about the way these components are arranged.
 
Last edited by a moderator:
Was that you that made that schematic Sleddog? I got it from this site, just couldn't remember off what thread.
 
Nice trace plot there, Monkey.
Very informative.

... Blue is AC probe from the power bus at fuse box (only shows the variance around the 13.5v DC it was idling at).

...blue shows the ignition events (including dwell time when voltage starts sinking)

... Powerbus voltage variance is 0.21v without any noticeable spikes...

Okay, the next test would be to probe the power line at or near the coil feed.
(Away from the dampening effect of the battery)
(That's also where most folks get power for their electronic ignitions)

Watch the noise.
Then, turn "off" the ignition switch.
Look for any grand finale spikes.

When the ignition switch is turned off, the battery is disconnected, and whatever's going on with inductive devices will be trapped in the "now isolated" power bus. Depending on happenstance, the ignition switch could be turned off while a coil is pulling current, or not, or in the middle of a sparking event, or in the middle of a charging event,...etc.

One of the many ignition failure mysteries in here seems to occur on restart attempts. The poster reports that the bike ran fine, then turns it off, then it won't restart. I've been suspecting some form of isolated or "trapped" inductive kickback during the switch to "off".

Found one in the alternator rotor.
http://www.xs650.com/threads/alternator-rotor-inductive-kickback.42605/
But, that experiment was conducted in a non-running scenario.

A similar test could be done to determine if the ignition coils do the same thing in a "running scenario"...
 
Last edited:
Back
Top