Getting HV spikes from LM317.

[HellTriX]

The title is a bit misleading but its somewhat true.

I have an LM317 set at 10V output. I am getting spikes to over 30V and 66v. I am even getting some negative 20-60V.

Now for what I'm doing in these circumstances. Firstly the input of the LM317 is bypassed with a 0.1uF cap, and the output with 1.0uF right at the component.

The output of the LM317 is running a 555 PWM circuit and is operating normally.
Also running from the output of the Vreg is a PWM driver chip. I put a 1.4k ohm resistor followed by a diode to the voltage input of the PWM driver to block any reverse voltage spikes since the PWM chip has a 330uF cap and a few smaller ones right at the component to serve as buffer (fastest switching) for the PWM output.

At low PWM signals the voltage from the Vreg is 10V, as the signal gets close to 50% pwm, the voltage would spike to over 30V. This is when I added the resistor and diode between the Vreg and the PWM driver. This blocked the positive voltage spikes.

But... At PWM signals greater than about 55% I start getting negative voltage readings, something like -10V to -55V. I am not sure what is causing this, but now that I think about it. I don't thing the circuit would be operating if this was really the case. So the more I think about it the more
I am questioning the accuracy of my trusty DMM.

What should I do to power my 555 with regulated voltage yet also provide regulated power to my PWM driver? I haven' yet to exceed 100mA drive current to the PWM driver chip, the most I have seen before the voltage goes all crazy is 80mA.

tnx

 

[MikeMl]

Can you post a schematic? A picture is worth a thousand words...

 

[colin55]

Are you measuring the circuit with a digital multimeter?

 

[HellTriX]

Are you measuring the circuit with a digital multimeter?

yep, all my analogs finally gave up after many many years of good use.


Also I have no schematic at this time. Was just hoping i could get some insight
as to possible causes. I'm guessing its just some sort of high frequency noise interfering with the DMM. I will try another meter when I get some time. Starting another long hard week of university math classes so don't have time to draw up a schematic.

tnx anyways.

 

[colin55]

Correct:

I'm guessing its just some sort of high frequency noise interfering with the DMM

Colin

 

[HellTriX]

I just swapped meters. Used a $2.99 harbor freight digital multimeter for the voltage and it was rock stable at 10.8V the entire duration of 0-100% PWM signal. I used the other DMM with the funny readings to measure the current, and around the same % of PWM the current reading nearly doubled which was just an error in the reading. So I guess some sort of harmonic of my PWM signal and the DMM just don't get along.

At least now this is resolved I can finish up the circuit, i was just worried about making magic smoke parts.

 

[Mikebits]

You need a scope to make an accurate diagnosis. Meters often give ambiguous results for quick transients.

 

[crutschow]

Also I have no schematic at this time..

How do you build a circuit without a schematic?

 

[HellTriX]

How do you build a circuit without a schematic?

You mean this isn't normal? lol
I have always done this. One component at a time and just build circuits
from my head. For each component there is a general idea of what it does, then you take into account its power requirements, frequencies, etc. And you add the required components for each chip added and such. Eventually you have a fully working circuit that in my case needs a few measurements and fine tuning and its complete.

The only time I have ever had to make schematics is when I wanted to ask a few questions about a circuit I had, and nobody will ever offer any information unless they see a schematic.

 

[BrownOut]

I'd say your readings were a fluke ( pun intended ) Had those been real spikes, you're IC would have fried.

 

[HellTriX]

I'd say your readings were a fluke ( pun intended ) Had those been real spikes, you're IC would have fried.

Thats where the majority of my confusion came from. I'm like this really cant be happening because that IC would have blown already. I guess I was just a bit worried because I have like $50 in mosfets in the circuit that I didn't want to destroy.

Got the circuit working properly last night after checking with other meters. I'm gonna move on to a high voltage test this week now. The test last night was at 14Vdc. Next test will be at 110V rectified and filtered AC to test that the control circuits stay at 10v and the mosfets do the same thing they did at 14v but at 110-150v.

 

[Hero999]

You need to test it using an oscilloscope to be sure.

If you can't post a schematic, please post a picture of your set up.

What does this contraption you're designing actually do? And what's it for?

 

[HellTriX]

Sure I can get a picture in a few hrs. its a PWM motor controller. Peak handling 250V at 1250 amps. Actual usage is 144V at 350-550A.
I have a scope, I will check the output tonight also.
I originally had the scope on the output to monitor the mosfet bank to make sure the switching times where sharp and low noise.

I will put the scope on the low voltage circuity to verify the voltage and signal levels. The only problem with my scope is it is limited to 10Mhz so if I have ringing above that I would never know.

 

[Hero999]

I doubt your DVM will show 10MHz signals.

 

[Noggin]

I think you should put a 1-10uF cap at the input of the LM317, and I can't imagine why you'd need a 330uF cap at the power rail of the 555. You'd be better off with a 1uF and a 0.1uF ceramic cap there instead.

You mean this isn't normal?

No, it's fairly normal. However, people that live and breath schematics have learned a few things...
1. Schematics are MUCH more efficient at expressing thoughts about a circuit than a human readable netlist.
2. Schematics are insanely useful in reproducing circuits
3. Schematics save time

 

[HellTriX]

I doubt your DVM will show 10MHz signals.

My Oscilloscope is only 10Mhz.

I think you should put a 1-10uF cap at the input of the LM317, and I can't imagine why you'd need a 330uF cap at the power rail of the 555. You'd be better off with a 1uF and a 0.1uF ceramic cap there instead.

I have a 1uf and .1uf at the 555.
The 330uF was at the PWM driver. Which has since been reduced to 6.7uF.
It is driving 10 parallel mosfet which have a gate capacitance of 53nC typical. Each mosfet has a 4.7 ohm gate resistor. PWM driver is a 14 amp driver so if I calculated it right, I used an order of magnitude greater buffer cap to the PWM driver input (10x). This latest change really reduced the noise of the circuit.

Its getting close but I literally have only 10 minutes here and there to poke at it and my weekend is already booked so still finding time to work on it or even a schematic is limited. I will be producing a schematic though at some point either when I get the circuit right, or find that I am stuck on something.

 

[Hero999]

My point is that you'll be able to see higher frequency signals on your 'scope than using a DVM.

What frequency is the 555 running at?

The maximum frequency you can get from a 555 is 500kHz so your 10MHz 'scope will be able to see the 20th harmonic which should be more than good enough.

 

[Noggin]

I have a 1uf and .1uf at the 555.
The 330uF was at the PWM driver. Which has since been reduced to 6.7uF.
It is driving 10 parallel mosfet which have a gate capacitance of 53nC typical. Each mosfet has a 4.7 ohm gate resistor. PWM driver is a 14 amp driver so if I calculated it right, I used an order of magnitude greater buffer cap to the PWM driver input (10x). This latest change really reduced the noise of the circuit.

Its getting close but I literally have only 10 minutes here and there to poke at it and my weekend is already booked so still finding time to work on it or even a schematic is limited. I will be producing a schematic though at some point either when I get the circuit right, or find that I am stuck on something.

Ok I gotcha now. Would have gotcha earlier if you had posted a schematic

(just bustin your balls)

 

[HellTriX]

What frequency is the 555 running at?

Uh. I think its somewhere between 12k and 15k. Been a while since I checked the frequency, but its audible and sounds normal.

Haven't had time to work on it in last few days. Also got a math exam (take home style) that I gotta work on.

 

[Hero999]

That should be no problem for a 10MHz 'scope.