Current spikes

dev/null

Hi all,

I have a robot with 4 battery packs, each with 7.2V that I connect to my stamps and circuitry.

I have lots of Stamps and circuitry connected to them, and they're all common grounded.

The problem is when I turn on the processors, I get voltage spikes which sends my servos to kingdom come positions, and my IC's behave like crazy (a shift register turn all outputs high). Sometimes they get stuck in a configuration until I restart the stamps.

Is there any way to prevent these spikes from entering my system?

Thanks.

Prettybird

Lots of capacitors mabe. Put 110ufd and .1 ufd to all battery inputs and depending on what kind other inputs. It may take come expermintation. Don't know your setup but that is a start.

dev/null

That's what I was expecting. Should I put them in series on the + terminal?

Like in the attachement?

Mike Green

No, filter capacitors like this case have to be in parallel with the battery. Make sure to connect the positive terminal of the electrolytic capacitor to the positive terminal of the battery. The smaller ceramic capacitors don't have a polarity.

Leon

TVSs are useful in that sort of situation.

Leon

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Amateur radio callsign: G1HSM
Suzuki SV1000S motorcycle

dev/null

I tried with 100uF and 0.1uF in parallell on the input lines, and the spikes are gone!!!

Thanks for help.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
---
dev/null.

dev/null

What's a TVS?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
---
dev/null.

Leon

Transient voltage suppressor. I use AVX TRansGuards.

Leon

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Amateur radio callsign: G1HSM
Suzuki SV1000S motorcycle

Alan Bradford

My guess would be that the I/O pins are glitchi'n during the Stamp(s) power up.
I have seen this happen on some of my industrial projects.
The proper procedure would be to power up the Stamp(s), initialize the I/O Pins then power up the Servos.
A .1 second delay is plenty if time for the Stamp to power up.
This shopuld keep things calmed down.





▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Alan Bradford ·N1YMQ

Plasma Technologies
Canaan NH 03741
www.plasmatechnologies.com

dev/null

It wasn't only the servos that were reacting. I have several IC's connected and they all went crazy.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
will a black hole evaporate if you send it to /dev/null ?

Alan Bradford

Im not sure what your device is.
On my industrial controllers,·the Stamp is powered as soon as the main power is applied.
I set all the I/O pins first then turn on one output that allows the rest of the circuitry to turn on.
That is all that runs until you press the Power On button.
This then energizes the rest of the system.
The Stamp cannot turn on the rest of the system, but it allows the Power On switch to do that.

That way the Stamp is in control before the rest of the system.
I do it with 24 Volt Relays and SSR's.
I am not concerned with power consumption or size.
The Plasma Cutters I control are sometimes as big as·40 ·by 200 feet, with 3 phase power.

Your design may need a diffrent type of·delay circuit to to the 2 step turn on.
Also if you have some IC's that also go nuts on power up, have them turn on early and then energize the external stuff they control.

Good Luck

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Alan Bradford ·N1YMQ

Plasma Technologies
Canaan NH 03741
www.plasmatechnologies.com

Alan Bradford

Hi Again,

I forgot you said it was a robot.
If the shift registers do not·see any garbage on the data lines they may behave properly.
With all of your processors, if you have a spare output available, you coult use it to reset the shift registers after the processors complete their power up.

Another thought...If you are powering up the whole system at once, some of the problems may be in the power supply.
It may be noisy before the regulators get stabilized.
Can you turn on the power section (and processors)·with a Master On switch, then have a second switch (or circuit) to energize the rest of the system, after a 100 ms delay?



▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Alan Bradford ·N1YMQ

Plasma Technologies
Canaan NH 03741
www.plasmatechnologies.com

dev/null

Yes. I think the proper way to go is to use a relay for all the circuitry. Problem is, I have 3 processors and I'm running out of pins! [noparse]:)[/noparse]
I have an 8 bit shift register, but I could get another one and control the relays with that.
Thanks!!!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
will a black hole evaporate if you send it to /dev/null ?

Sandgroper

Are the logic circuits and power supplies for the IC's separate from the load circuits? If not, dual power supplies can be useful. 7.2v can get very close to a chip's threshold voltage if there is significant current draw or if the batteries aren't fully charged.

Keeping the lines short can help to reduce noise too.

dev/null

I have four 7.2V battery packs on the robot. They are 4200mAH.
- Pack one powers two BS2p's and one Javelin, including the relays, IC's and peripherals connected to them.
- Pack two powers a PSC, connected to the relay.
- Pack three powers the motors to the wheels, connected to the relay.
- Pack four powers a headlight, connected to the relay.

I'm going to have to get another 12V pack to power my relays, as they operate on 12V. They will work on a 7.2V pack for a while but quickly goes dead at about 6.5V. I have another 4 relays I'm going to connect as well to get more control over power to the individual servos (curently 8 servos).

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
/dev/null

dev/null

Here are some pictures of my robot, R2B2. WORK IN PROGRESS!

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
/dev/null