Back EMF smoke in controll panel

[john-100]

Hi Tom and mirocha


parasitic oscillation a possibility especially if the connecting wires are long

0.1 uf capacitors across the input and output terminals and the TIP147's base collector terminals could help



I thought I find I am going back to is the original current dumping circuit would have been helped
by the other axis drivers taking some of the returned energy


with this version only being connected to one axis



the 33 ohm resistor could be dissipating 175W when the TIP147 switches on

if it takes too long to switch off the 33R 10W resistor is going to have a short life


the other possibility is the heat sink is too small
and the TIP 147 over heating when the base current drops and is no longer saturated

when the collector current is about 1A and the collector emitter voltage is 35V
you need to have a heat sink big enough to dissipate over 35W ! ( so will the 33 ohm resistor )



John


PS
unknown Chinese TIP147's could be rejects or fakes - cheaper transistors that's been re-marked

[mirocha]

Hi Tom and mirocha


parasitic oscillation a possibility especially if the connecting wires are long

0.1 uf capacitors across the input and output terminals and the TIP147's base collector terminals could help



I thought I find I am going back to is the original current dumping circuit would have been helped
by the other axis drivers taking some of the returned energy


with this version only being connected to one axis



the 33 ohm resistor could be dissipating 175W when the TIP147 switches on

if it takes too long to switch off the 33R 10W resistor is going to have a short life


the other possibility is the heat sink is too small
and the TIP 147 over heating when the base current drops and is no longer saturated

when the collector current is about 1A and the collector emitter voltage is 35V
you need to have a heat sink big enough to dissipate over 35W ! ( so will the 33 ohm resistor )



John


PS
unknown Chinese TIP147's could be rejects or fakes - cheaper transistors that's been re-marked

Just talked with the founder owner of Gecko, he walked me through every thing. The original circuit with adding a 6.8k resistor and an LED will add about 10 milliamps and the circuit will shunt about 2.18 amps so the TIP 147 at 10 amps is more than adequate. The problem was my originally using the 1000uf Capacitor which drew enormous current.

We did calculations on everything, running a single stepper motor return energy could raise the voltage on the 70 volt power supply as much as 10 volts. Running several stepper motors, the other motors will absorb return energy.

I will build the original circuit adding just the 6.8k resistor and an LED using a quality TIP 147, no more Chinese crap.

I think this the way for me to go. Thanks for every ones help and input.

I will post how it works out once I get the new transistors.

Thanks again

Joseph Mirocha

[mirocha]

Just talked with the founder owner of Gecko, he walked me through every thing. The original circuit with adding a 6.8k resistor and an LED will add about 10 milliamps and the circuit will shunt about 2.18 amps so the TIP 147 at 10 amps is more than adequate. The problem was my originally using the 1000uf Capacitor which drew enormous current.

We did calculations on everything, running a single stepper motor return energy could raise the voltage on the 70 volt power supply as much as 10 volts. Running several stepper motors, the other motors will absorb return energy.

I will build the original circuit adding just the 6.8k resistor and an LED using a quality TIP 147, no more Chinese crap.

I think this the way for me to go. Thanks for every ones help and input.

I will post how it works out once I get the new transistors.

Thanks again

Joseph Mirocha

Installed new transistors, using the original gecko return energy dump circuit only adding an LED and 689k resistor and the circuit worked with no problem. The LED is off and when the stepper motor quickly slows down the LED turns on briefly letting me know the return energy dump circuit is working.

New problem, each drive has a separate return energy dump circuit and a separate LED for each for each return energy dump circuit. I extended with wire and mounted the LED to the control cabinet door. When the LED is mounted directly to the circuit board the LED works perfect when mounted to the cabinet door by wire the LED is always glowing not in the off state.

The wire length from the circuit board to the each of the LEDs is between approximately 30" and 46". The LED with the longer wire leads flows the brightest and the LED with the shorter wire leads glowing significantly less.

I mounted terminal blocks in the circuit board to connect the wire leads to the LED. If I connect the LED directly on the circuit board the LED is off and turns on when there is return energy as it should, when I connect the LED with 30" or longer wire leads the LED is always glowing and flashes brighter when there is return energy.

I also swapped the return energy circuits from one drive to the other to see if the problem moved to see if it was the circuit board causing the problem. Board one with the longest wire leads to the LED had the brightest glowing LED, I swapped it with board four which had the shortest wire leads and the problem did not move so it appears the problem is the wire, I don't know what to do to fix it, just be capacitance or inductance problem I would think.

Asking for help in how to solve this issue.

Thanks joe

[TomKerekes]

Hi Joseph,

I still suggest adding bypass capacitors to your circuit. In my experience a high speed switching circuit like that will often do bizarre things without them. Those big capacitors will not act like capacitors at all at high frequencies.

Let us know what you find.

Good luck and Regards

[mirocha]

Hi Joseph,

I still suggest adding bypass capacitors to your circuit. In my experience a high speed switching circuit like that will often do bizarre things without them. Those big capacitors will not act like capacitors at all at high frequencies.

Let us know what you find.

Good luck and Regards

I ordered 30 - .1uf 160 volt, ceramic capacitors from digikey, where do you recommend installing them, output, input, across output to LED, across transistor base and collector any of those places or all those plsces?

[john-100]

if your using the circuit from post 20 this is what I would try -



John

[TomKerekes]

Here are my thoughts:

I would put one as shown at the output across C1 to suppress high frequency coming from Geckos (can't hurt)

I think one should be placed on the supply input. I'd put this as most important to have a stable solid supply voltage for the circuit.

Not sure the one across the LED would be helpful (but wouldn't hurt)

The other two on the Transistor I'm not sure about. A voltage spike on the Output could send an unlimited current through the Base-Emitter of the transistor. I think it would slow down the turn on/off time of the transistor causing it to dissipate more heat.

But a lot of this is like black magic to me

[mirocha]

Here are my thoughts:

I would put one as shown at the output across C1 to suppress high frequency coming from Geckos (can't hurt)

I think one should be placed on the supply input. I'd put this as most important to have a stable solid supply voltage for the circuit.

Not sure the one across the LED would be helpful (but wouldn't hurt)

The other two on the Transistor I'm not sure about. A voltage spike on the Output could send an unlimited current through the Base-Emitter of the transistor. I think it would slow down the turn on/off time of the transistor causing it to dissipate more heat.

But a lot of this is like black magic to me

I will also add a 470uf capacitor within 1" of the gecko drive to the power in on the gecko drive gecko states this must be done if the supply is more than 1 foot from the drive, this something I missed and did not do.

I'll post my results just have to wait for parts.

[TomKerekes]

Hi Guys,

Bypass capacitors have always been somewhat of a mystery to me so I thought I'd do a little research. All I know is I often have trouble if I don't add them. Here are a few bits of info I found that I thought were interesting. Hopefully I'm not hijacking your thread too much but I think it might be related.

I found a few references on the inductance of a loop of wire such as:
Estimating wire & loop inductance: Rule of Thumb #15 | EDN

The rule-of-thumb seems to be around 1uH per meter of wire.

So for example if your wiring from your supply to your clamp circuit is 20inches (each way) then the inductance in the wiring would be about 1uH.

The formula of impedance, X, of an inductor at a given frequency, F, is:
X = 2 Pi F L

For sake of discussion an oscillation at 1MHz would see ~6 ohms

A current of +/-1 amp at 1 MHz would cause the supply voltage to change +/- 6V !!

It only takes the Supply Voltage to drop by ~0.7V for the circuit to turn on the clamping transistor.


I also found this article on Ceramic vs Electrolytic Caps. They don't seem to be as much different as I expected. From this eetimes article
Power Tip 51: Be aware of capacitor parasitics | EE Times

There is this chart:


At 1MHz the ceramic capacitor has an effective resistance of ~0.008 ohms
At 1MHz the Electrolytic has an effective resistance of ~0.03 ohms

Either one would be much better than the 6 ohms without any capacitor.

Actually because of the bigger value the Electrolytic is more effective than the Ceramic until 300KHz

And that chart is for a 22uF Ceramic. The 0.1uF I recommended still acts properly as a capacitor beyond 1MHz but its low value makes its impedance ~1 ohm at 1MHz. So a big Electrolytic would be more effective. The 0.1uF Ceramic is more appropriate for killing 10MHz oscillations. Here is a plot from an AVX 0.1uF Ceramic Capacitor:



Its hard to say if the circuit is breaking into oscillation what the frequency would be without scoping it. I believe the standard approach to cover the entire frequency band is to use a big Electrolytic + 10uF Tantalum + 0.1uF Ceramic.

I'm not sure I'm being of much help

Regards

[RCaffin]

It's a bit more complex than that, but no matter.
Bypass caps on power INputs to drivers: yes indeed.
Bypas caps on the outputs of switch-mode drivers to servos and steppers: NO. Very bad idea. Driver may stop working.

Cheers
Roger

[mirocha]

Hi Guys,

Bypass capacitors have always been somewhat of a mystery to me so I thought I'd do a little research. All I know is I often have trouble if I don't add them. Here are a few bits of info I found that I thought were interesting. Hopefully I'm not hijacking your thread too much but I think it might be related.

I found a few references on the inductance of a loop of wire such as:
Estimating wire & loop inductance: Rule of Thumb #15 | EDN

The rule-of-thumb seems to be around 1uH per meter of wire.

So for example if your wiring from your supply to your clamp circuit is 20inches (each way) then the inductance in the wiring would be about 1uH.

The formula of impedance, X, of an inductor at a given frequency, F, is:
X = 2 Pi F L

For sake of discussion an oscillation at 1MHz would see ~6 ohms

A current of +/-1 amp at 1 MHz would cause the supply voltage to change +/- 6V !!

It only takes the Supply Voltage to drop by ~0.7V for the circuit to turn on the clamping transistor.


I also found this article on Ceramic vs Electrolytic Caps. They don't seem to be as much different as I expected. From this eetimes article
Power Tip 51: Be aware of capacitor parasitics | EE Times

There is this chart:


At 1MHz the ceramic capacitor has an effective resistance of ~0.008 ohms
At 1MHz the Electrolytic has an effective resistance of ~0.03 ohms

Either one would be much better than the 6 ohms without any capacitor.

Actually because of the bigger value the Electrolytic is more effective than the Ceramic until 300KHz

And that chart is for a 22uF Ceramic. The 0.1uF I recommended still acts properly as a capacitor beyond 1MHz but its low value makes its impedance ~1 ohm at 1MHz. So a big Electrolytic would be more effective. The 0.1uF Ceramic is more appropriate for killing 10MHz oscillations. Here is a plot from an AVX 0.1uF Ceramic Capacitor:



Its hard to say if the circuit is breaking into oscillation what the frequency would be without scoping it. I believe the standard approach to cover the entire frequency band is to use a big Electrolytic + 10uF Tantalum + 0.1uF Ceramic.

I'm not sure I'm being of much help

Regards

Ok so I'm feeding 70 volts to the return energy circuit which is feeding the gecko drive and the return energy dump circuit also has 30" or a little longer wire leads feeding an LED.

I am adding a 470uf capacitor directly on the gecko drive voltage input, the output of the gecko drive goes directly to the stepper motors with no capacitors.

The return energy dump circuit is feeding the gecko drive. The return energy dump circuit has a 1000uf capacitor across the output that hat is feeding the gecko drive input which has a 470uf capacitor across it.

Capacitors are cheap so to cover the entire frequency band what would be the best capacitors to add. If there is a better value than the .1uf ceramic i can order whatever would be best.


I am assuming you are saying to parallel the capacitors to cover the entire range.

Should I put a .1uf ceramic and a 10uf tantalum capacitor in parallel across both the input and output of the return energy dump circuit and do I need capacitors across the wire leads to the LED, it seems the long wired are causing the problem or will the input and\ or output capacitors going to solve the problem?

So what would be the best size capacitors to use and where to put them in the return energy dump circuit?

Don't be concerned with hijacking the thread I looking for help, thanks.

Joe

[RCaffin]

You are probably massively over-thinking al this. The big electrolytics are fine, but That is all I use, and I don't bother with an energy dump circuit at all. I am doing semi-production machining.

Cheers
Roger

[mirocha]

You are probably massively over-thinking al this. The big electrolytics are fine, but That is all I use, and I don't bother with an energy dump circuit at all. I am doing semi-production machining.

Cheers
Roger

Originally I was using regulated power supplies and the drives failed and locked out without the return energy dump circuit.

I know the return energy dump works, I found that adding just a resistor and a diode mounted on the circuit board it worked with no problem and the diode let's me see every time the motors are returning energy back into the other motors and the power supply. I was told by the gecko engineer that return energy returned could be as high as 10 volts.

Now I have changed to an unregulated power and the return energy dump circuit may not be needed, however I see it as a nice option, kind of like a power on indicating lamp, you don't need it but nice to know if the power is on.

The return energy circuit works and it works with the LED mounted on the circuit board it's when I add 30" of wire to the LED and mount the LED to cabinet door that I get a constant glow in the LED however the LED flashes brightly for every return of energy from the motor.

Why pump a higher voltage back into the system if you don't need too, with LED letting you know when the motors are regenerating i could test and see how much I would need to reduce deceleration rate to reduce or eliminate return energy from the motors.

Not so much that I am over thinking but looking to add an option that is functional letting me know what my system is doing, I also plan to mount an amp/volt meter just so I can see the actual not calculated max current with all motors running under load, kind of like a power on light not needed but a nice option. If the motors are moving you know the power is on why have a power on light.

I am so close, the return energy dump circuit works, the LED on the circuit board works, just need to get the LED to work when fed with 30" of wire remote from the circuit board.

I think so far for a 65 year old guy with no electronics schooling I'm doing pretty good.

I have kflop working, kanalog working, kmotion working and stepper motors turning using gecko drives at 10x microstepping. This is cool stuff.


Thank you for your past and hopefully future comments and suggestions.

Joe

[TomKerekes]

Hi Joe,

Ok so I'm feeding 70 volts to the return energy circuit which is feeding the gecko drive and the return energy dump circuit also has 30" or a little longer wire leads feeding an LED

I'm curious how long are the wires from the Supply to the "Return Energy Dump board". Also the distance from the circuit to the Gecko.

I'm guessing with the 30 inch LED connection, one of the wires which is connected to your Circuit GND is picking up noise from the stepper motors and introducing it into the dump circuit GND. Either that or its parasitic capacitance to Earth GND is affecting the circuit. It might be fun to run a test. Connect the LED back on the board and verify the circuit works properly. Then add a 30 inch wire like an antenna to the GND of the dump circuit. Then see if the circuit behaves weird.

Capacitors are cheap so to cover the entire frequency band what would be the best capacitors to add. If there is a better value than the .1uf ceramic i can order whatever would be best.

Hmmm Normally I'd put a 10uF Tantalum. But I see a 100V version would be prohibitively expensive. A 10uF ceramic is pricey to. Maybe like a 100V 2uF Ceramic. Like this for $0.65:

https://www.digikey.com/product-deta...9-1-ND/5812064

It has a resistance below 0.1ohm at 1 MHz see the chart

I am assuming you are saying to parallel the capacitors to cover the entire range.

Yes

Should I put a .1uf ceramic and a 10uf tantalum capacitor in parallel across both the input and output of the return energy dump circuit

Yes, or the 2uF Ceramic if 10uF Tantalum is too expensive.

do I need capacitors across the wire leads to the LED, it seems the long wired are causing the problem or will the input and\ or output capacitors going to solve the problem?

Correct, the idea is if the supply voltages on the circuit board are stable then the circuit should behave properly. If the theory is right that noise is coming in the LED wire connected to your circuit GND, then adding a capacitor across the LED (at either end) wouldn't help. Actually now that I think about it it could hurt. Because any noise coming in the other LED wire to high resistances can't really effect anything. But adding a cap would couple it's noise into your GND as well and add to the problem. (a cap is like a short circuit to high frequency noise).

So what would be the best size capacitors to use and where to put them in the return energy dump circuit?

So again I'm guessing that pretty much any capacitor on the Supply Input may solve the problem. But to cover all theoretical bases 3 in parallel on the Input. There is already the 1000uF on the output. Adding 2 more smaller caps to handle higher frequencies wouldn't hurt.

Some of us like to over think things

I did the math for a medium sized size 23 stepper (moment of inertia 8.4e-5 Kgm^2) running at 1000RPM has a kinetic energy of 0.45Joules. That energy would theoretically raise a 1000uF Capacitor from 70V to 77V.

Also for a 50Kg Gantry moving at 600ipm has an energy of 1.5Joules. That energy would raise a 1000uF Capacitor from 70V to 89V

A long lead screw or large diameter pulley might have even more energy.

There is a fact that surprised me when I first realized it and I often ask it on Job interviews. If any object is spinning at some rate and has a certain amount of kinetic energy, if you were to double the "size" of the object and spin it at the same rate, how many times more energy would it have? For example a 1 inch cube spinning at 1000RPM vs 2 inch cube spinning at 1000RPM. Anyone know the answer?

Regards
TK

[RCaffin]

You can not only over-think some things, you can also over-engineer them. Yep, really.

The circuitry inside a modern driver, such as Gecko 201X or 320, has a lot of very high frequency stuff. The designs for all of these ASSUMES a simple diode bridge PS. If you use a Switch Mode PS you risk injecting the 'noise' or RFI from the PS into the driver - and totally stuffing the whole thing. Using an SMPS on some of these drivers will actually void the warranty! It's that serious.


I use a large transformer, a diode bridge, a large capacitor at the bridge, then some heavy twin-flex to where the drivers are (a different box), and another smaller cap there. Then power goes to each driver via a separate cable. Before the power gets there it goes through a 10A meter and a 5 A magnetic (resettable, fast) circuit breaker. There may be a 0.1 uF foil cap right at the input terminals. This is for semi-production use.

I do not have any energy dump systems on the axis drivers. I do have an energy dump on the spindle driver relays which is brought into operation with the M5 command. But that is all.

Cheers
Roger

[mirocha]

Hi Joe,

I'm curious how long are the wires from the Supply to the "Return Energy Dump board". Also the distance from the circuit to the Gecko.

I'm guessing with the 30 inch LED connection, one of the wires which is connected to your Circuit GND is picking up noise from the stepper motors and introducing it into the dump circuit GND. Either that or its parasitic capacitance to Earth GND is affecting the circuit. It might be fun to run a test. Connect the LED back on the board and verify the circuit works properly. Then add a 30 inch wire like an antenna to the GND of the dump circuit. Then see if the circuit behaves weird.

Hmmm Normally I'd put a 10uF Tantalum. But I see a 100V version would be prohibitively expensive. A 10uF ceramic is pricey to. Maybe like a 100V 2uF Ceramic. Like this for $0.65:

https://www.digikey.com/product-deta...9-1-ND/5812064

It has a resistance below 0.1ohm at 1 MHz see the chart



Yes

Yes, or the 2uF Ceramic if 10uF Tantalum is too expensive.

Correct, the idea is if the supply voltages on the circuit board are stable then the circuit should behave properly. If the theory is right that noise is coming in the LED wire connected to your circuit GND, then adding a capacitor across the LED (at either end) wouldn't help. Actually now that I think about it it could hurt. Because any noise coming in the other LED wire to high resistances can't really effect anything. But adding a cap would couple it's noise into your GND as well and add to the problem. (a cap is like a short circuit to high frequency noise).

So again I'm guessing that pretty much any capacitor on the Supply Input may solve the problem. But to cover all theoretical bases 3 in parallel on the Input. There is already the 1000uF on the output. Adding 2 more smaller caps to handle higher frequencies wouldn't hurt.

Some of us like to over think things

I did the math for a medium sized size 23 stepper (moment of inertia 8.4e-5 Kgm^2) running at 1000RPM has a kinetic energy of 0.45Joules. That energy would theoretically raise a 1000uF Capacitor from 70V to 77V.

Also for a 50Kg Gantry moving at 600ipm has an energy of 1.5Joules. That energy would raise a 1000uF Capacitor from 70V to 89V

A long lead screw or large diameter pulley might have even more energy.

There is a fact that surprised me when I first realized it and I often ask it on Job interviews. If any object is spinning at some rate and has a certain amount of kinetic energy, if you were to double the "size" of the object and spin it at the same rate, how many times more energy would it have? For example a 1 inch cube spinning at 1000RPM vs 2 inch cube spinning at 1000RPM. Anyone know the answer?

Regards
TK

The wires from the power supply to the first return energy dump circuit are about 10" and get progressively longer to the last return energy dump board to about 24" long to the last return energy dump board.

all the wires from the return energy dump circuit board to the gecko drives are about 6" long.

the wires from the return energy circuit board to the LED with the longest wire is about 44" which is the return energy circuit board with the shortest power supply wires and is the one that has the brightest glowing LED.

when power is applied to the gecko drives the LEDs glow with the LED with the shortest wires glowing the least with each LED getting brighter as the wires are longer. This with the gecko drives on holding the motors still.

I made the energy dump circuit boards with terminals blocks to wire the LED so it was easy to do your test.

With an LED connected directly to the return energy dump circuit board and the gecko drive powered on the LED was off not lit. I had disconnected the wires to the LED that were 44" long. I touched one of the disconnected LED wires to the ground connection of the LED on the return energy dump circuit board, it did nothing. As a second test I touched one of the disconnected LED wires to the positive terminal of the LED mounted on the return energy circuit board and the LED on the return energy dump circuit board glowed, I did it numerous times, every time I touched the wire to the positive the LED would glow. Your test proved that the wire is acting like an antenna.

The motors I will be using will be nema 34 at 1260 oz/in., The motors I am testing with are nema 34 at 1600 oz/in.

When I tried a 1000uf capacitor across the LED with a diode to keep the LED lit longer it was to much instantaneous current and shorted the return energy dump transistor so I need to be careful not to short out the transistor with anything across the LED. I agree not to put anything across the LED at this time.

I agree that the 10uf tantalum cap is to pricy, your other suggestions was a 2 uf ceramic which is cost effective, I also see digikey has a 4.7uf ceramic for less than a dollar which would be better?

Thanks Joe
With the spinning cubes, you double the size did you double the weight?

[TomKerekes]

Hi Joe,

Thanks for clear descriptions.

I made the energy dump circuit boards with terminals blocks to wire the LED so it was easy to do your test.

With an LED connected directly to the return energy dump circuit board and the gecko drive powered on the LED was off not lit. I had disconnected the wires to the LED that were 44" long. I touched one of the disconnected LED wires to the ground connection of the LED on the return energy dump circuit board, it did nothing. As a second test I touched one of the disconnected LED wires to the positive terminal of the LED mounted on the return energy circuit board and the LED on the return energy dump circuit board glowed, I did it numerous times, every time I touched the wire to the positive the LED would glow. Your test proved that the wire is acting like an antenna.

Hmmm that pretty much blows my theory. I'm surprised a 44inch piece of wire can pick up (or radiate) enough energy to light an LED. I'm curious and certainly open to suggestions what is going on. I suppose shielded cables to the LED would solve the problem. Are the LED wires bundled together or close to other things like Stepper Motor wiring or other Gecko Drivers Supply wiring?

With the spinning cubes, you double the size did you double the weight?

The 2 inch cube is more than double the weight than the 1 inch cube. Hint: the answer is surprisingly big

Regards

[john-100]

having seen the results of tantalum capacitors setting fire to the PCB in an ASTON 4 character generator ( used a TV station's graphics department )

I would use solid aluminium electrolytic capacitors instead of tantalum capacitors

Is the Gecko drivers power supply negative connected to mains supply earth / ground ??
assuming you are using a metal enclosure do you have a braided strap connecting the door to the enclosure ??

John

[mirocha]

Hi Joe,

Thanks for clear descriptions.

Hmmm that pretty much blows my theory. I'm surprised a 44inch piece of wire can pick up (or radiate) enough energy to light an LED. I'm curious and certainly open to suggestions what is going on. I suppose shielded cables to the LED would solve the problem. Are the LED wires bundled together or close to other things like Stepper Motor wiring or other Gecko Drivers Supply wiring?

The 2 inch cube is more than double the weight than the 1 inch cube. Hint: the answer is surprisingly big

Regards

I did some more testing, using the old method of smell and touch. I found one of the 4 gecko drives was running hot, I had a gecko G201(same as a G210 but with out the half step option board) and swapped it with the gecko that was running hot. Now instead of 4 LEDs lit connected with the wire mounted on the cabinet door I had one LED lit. I moved the gecko drive that had the LED lit to a different axis and the problem moved with it. I think I have 2 bad gecko drives that are feeding noise back into the system. One bad one that is running hot and one bad one that is sending noise into the system.

I am sending the two drives back to gecko for testing and repair.

I have a couple of thoughts about the return energy dump circuit. If I am correct and the bad drive that is not getting hot is sending noise into the system its a good thing that the LED was glowing telling me the drive is corrupting my system with noise, good feedback once you know what the glowing LED means. Also I know from failures that if the LED stays on the return energy dump circuit has failed. And if all is well the LED will flash on during quick deceleration letting you know the return energy is being shunted to ground.

I find it strange that when the LED was mounted to the circuit board and not on long wire leads that it appeared there was no problem. I am confidant that the two drives are bad one I could hardly touch it was so hot.

Now what do I do with the capacitors I ordered, do I still install them as a precautionary step. I now think that there was a lot of noise on those wires so would adding the capacitors give me a little protection or take away the ability to know if the drives are back feeding noise into the system by filtering the noise that was making the LEDs glow?

All in all I think this is a good circuit for knowing a little more on what is going on in the system. Just over thinking the whole thing.

Once I get a fourth drive that is either fixed or replaced I'll post results.

Tom let me know your thoughts if I'm back on the right track, and yes I did have a ground wire from the cabinet to the door and the negative of the power supply is case grounded.

Joe

[TomKerekes]

Hi Joe,

I think you are on the right track.

The Supply Bypass Caps I recommended may not be necessary. Sorry if I sent you on a wild goose chase. I tend to always add them if I can't convince myself they are not absolutely unnecessary.

What gauge wire is your supply wiring btw?

Regards