Spin 2 Low-Z testing

Marcos' unmaintained, but still in-use, Puma for FreeEMS circuit board/hardware design!
User avatar
nitrousnrg
LQFP144 - On Top Of The Game
Posts: 468
Joined: Tue Jun 24, 2008 5:31 pm

Re: Spin 2 Low-Z testing

Post by nitrousnrg »

See it yourself:

http://www.national.com/ds/LM/LM1949.pdf

page 10 :-)

It play a role in the frequency and look of the PWM, I see it as a configuration resistor. Current measuring R is 0.1ohm.

Oh, and second scope shot, its Vce, 5v/div.

You can see that the transistor has a Vce drop of nearly 5v at 2A, while the datasheet claims 2.5V max at a 3amp or so. Spikes are clamped at 34 volt or so.
Marcos
User avatar
jharvey
1N4001 - Signed up
Posts: 1607
Joined: Tue Jun 10, 2008 5:17 pm

Re: Spin 2 Low-Z testing

Post by jharvey »

Hmm, Vce issue, the LM claims it can provide 22ma, is it giving you enough drive? I might guess that's 22mA with a low voltage drop, your transistor might have a slightly higher junction voltage, causing less than 22mA from the LM. The base current is a bit harder to measure. Have you see the CA60? It could give you up to like 20khz AC down to DC, with out really modifying the circuit.

http://www.veejer.com/testequipment.html#mACC
User avatar
nitrousnrg
LQFP144 - On Top Of The Game
Posts: 468
Joined: Tue Jun 24, 2008 5:31 pm

Re: Spin 2 Low-Z testing

Post by nitrousnrg »

Thanks for the link, I saw it somewhere (you probably did show me it before).

Okay, the PWM is working, it was a just mix of a bridge and and dumb mistake.

The transistor saturates nicely, and holds at 1A. The injector never reaches the 4A peak, but I don't really care about that.

This is Vce, you can see the line at Vbat, then zero when the transistor saturates, and later it comes the PWM ringing from 0 to Vzener. The zener is 33v, althoug the PWM climbs up until 37v or so.
Image

And current graphs:

Image

The switching frequency is high, nearly 50khz. I played for a while with two configuration resistors that have influence over the switching frequency, but it doesn't seem to do anything.

According to the datashet,

octave:33> f = (Rs/L) * (RA/RB) * (VBATT/VZ) * (1 - VBATT/VZ)
f = 3841.19790278148 hertz

I changed RB from 2ohm to 12 ohm, and could't notice the difference, so something weird is happening in there. I always see 50khz. I'll check the behavior of the 2N6045 just to make sure is not a difference in the SMT darlington.

50 commutations per milisecond is a lot, and unnecessary given the mass and inductance of an injector, hence the efforts trying to make it slower. I'd like 10khz to 15khz. Keep in mind its a 37v excursion. Also, I find higher frequencies harder to filter, and lots of stuff that I don't like could appear everywhere.

Its cold outside (6°C) to do a temperature test. Probably because its 3:41 am in the winter. I like winter, but these things are noisy and I don't want to bother my neighbors with this, this late. Tomorrow I'l put the board working for a while to see the temperature rise of the transistor, and especially of the zener.

So, slow down the frequency, then check temperatures, then measure noise in the test board, in the mbed, at the battery, and in different grounds of the car. I believe noise won't be a problem having the dirty/clean ground setup, many folks do this out in the field.

Also, thanks jim for the mail, I read it but at that time I didn't try anything of this to write a worthy reply :-)
Marcos
User avatar
nitrousnrg
LQFP144 - On Top Of The Game
Posts: 468
Joined: Tue Jun 24, 2008 5:31 pm

Re: Spin 2 Low-Z testing

Post by nitrousnrg »

On second thought, 10khz could be too low. My only input is a very slow, very large injector, but if it would be a fast one, the current ripple could be too large.
Marcos
User avatar
jharvey
1N4001 - Signed up
Posts: 1607
Joined: Tue Jun 10, 2008 5:17 pm

Re: Spin 2 Low-Z testing

Post by jharvey »

I kind of wonder if adding an 80V supply and mimicking that TLE direct drive chip might help get to that 4 amps, or a faster on time.
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Spin 2 Low-Z testing

Post by Fred »

Marcos, re the high frequency, I wouldn't worry about current ripple, but what you should worry about is off time accuracy. With the HF the signal dies more or less as it dies, but with the LF it could die noticably sooner. IE, you will end your pulse upto half a cycle earlier than what you intend to. Thus, to maintain accurate PWs (which are of critical importance for typical low z applications where resistors are not acceptable) you must PWM it at a high rate.

Re the noise protection, I advise that you read the old thread on low z control and implement the transistor + zener noise catching mechanism, zeners capable of such current ahd power dissipation are not common, cheap, or high performance, transistors are. You want this to be reliable and low noise, right? :-)

Reality of PWM is heat, not as much as resistive, but still a significant amount which you must manage. Putting it into the PCB is not an acceptable approach in a compact design. No to220, ok, no problem, but you need another solution to get the heat out, there are many, design one! :-)

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
nitrousnrg
LQFP144 - On Top Of The Game
Posts: 468
Joined: Tue Jun 24, 2008 5:31 pm

Re: Spin 2 Low-Z testing

Post by nitrousnrg »

Jared, the injector has a datasheet that says around 12v for the supply. I'd say that with 80v it reaches the 4A :)


I got it, LF = not good.

I ran a test on the car, and with the original freq, ~50khz, the disturbance seen was mostly generated by the peak of current consumption, rather than the PWM. Measuring at the battery, I see the 12v fluctuating at the rythm of the injector. The uglyness comes from a non-filtered-yet switched power supply.

10mV/div:
Image

The heat isn't a concern in this case. It heats up, but it doesn't worry me. At a high duty%, and after several minutes working, it is hot, but not much (oh, language). Being always in hold and with a resistive load there isn't much temperature either. The hottest part can be touched.... I don't have experience on measuring temperature with my hands, but if I had I'd say how many degrees hot it is :-/

The zener dosen't get hot as I expected, and the transistor still has room for more dissipation (there are no vias in this pcb:-). I saw heatsinks that are soldered in the pcb plane. It seems like a cool way to get rid of the heat, no screws.
Marcos
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: Spin 2 Low-Z testing

Post by Fred »

nitrousnrg wrote:The heat isn't a concern in this case. It heats up, but it doesn't worry me. At a high duty%, and after several minutes working, it is hot, but not much (oh, language). Being always in hold and with a resistive load there isn't much temperature either. The hottest part can be touched.... I don't have experience on measuring temperature with my hands, but if I had I'd say how many degrees hot it is :-/
LOL, you're hot! (52.746 degrees in my experience) :-p
I saw heatsinks that are soldered in the pcb plane. It seems like a cool way to get rid of the heat, no screws.
I don't know that you did, unless they were made of copper or silver! They probably had steel mounting pins pressed into the allow, and the steel was soldered. I fixed a few dry joints on one like this in a stationary TV a few days ago (borrowed a soldering iron and was almost too happy about it), so vibration could be an issue, depending on ratio of heatsink to mount points and geometry.

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
User avatar
nitrousnrg
LQFP144 - On Top Of The Game
Posts: 468
Joined: Tue Jun 24, 2008 5:31 pm

Re: Spin 2 Low-Z testing

Post by nitrousnrg »

Fred wrote:LOL, you're hot! (52.746 degrees in my experience) :-p
ha! I'd say its around 50°C, but i have no clue really. I reacall your TA board was at that temperature.
so vibration could be an issue, depending on ratio of heatsink to mount points and geometry.
Yup, thats my big concern, but I saw one that was not tall, with a big surface, and full of little aluminum towers :)
Marcos
Post Reply