Amplified thermocouple issues

<< ---------------------- I'm on 640 x 480 ---------------------------------- >>
<< ---- i have to 'pan' left and right if its much wider than this ------- >>

Hi Ryan,

I've had a bit of a think,
and i would like you to try something.
I would like you to move the ground wire on that amplifier unit over
to the Thermocouple ground connection.
I am not sure about the fixings on the amplifier, but my guess is that
they are not internally connected to anything.
However, just to be sure, would you mind just temporarily removing
them and maybe putting a bit of cardboard under the amplifier,
maybe a bit of tape or something similar, it can be screwed back down
afterwards, i just want to be sure that there is only one ground track,
for this check.

The gauge, ... i agree, it doesn't look like it has an earth/ground
connection at all. Thats ok, it doesn't need to have.

The lamp in the gauge does look like it is not connected in any way to
the circuitry involved with it.

The unit is supposed to run through a 1 Ampere fuse, so it takes less
than an amp, i don't see a problem running it through the braid, thats
the shielding around the thermocouple signal wire.
But it might be interesting to see just how much current it takes.
But don't worry about the current for this check.

I've included a pic, partly copied from the links you gave me.
Doing this should give a similar situation as you had when you gave
it a separate supply.

Lets see if this gives some sensible results, i have my fingers
crossed !

Cheers, John

Been there already. That was one of my first attempted fixes, no go.
The ground to the light definately does not connect to the gauge in any way. The amp enclosure is all plastic and the terminal block has insulating foam on the bottom to prevent shorts. The only ground source/s coming in are from the TC and the ground lead.
Already did a current check last night. This threw me a curve and I'm still questioning that I did it right. On the 2mA scale on my MM, I get a reading between .001mA and .002mA. It's a digital meter so it just kept going between the two values. Is it just me or does that seem low? I know it's accurate, but geeze. I know it's not a conventional gauge so I wonder if it only really draws current when the needle is in motion? If that's the case then I'm not even sure how to get a good reading w/o data aquistion of some sort. I can torch the TC for quick movement, but watching a digital MM for changing values isn't really accurate.
I'm guessing however that 2W DC/DC converter would be plenty for it, it was rated at up to 167mA so I would presume that should do the trick,,,, unless of course another solution is somehow found - fingers crossed. Real puzzles are always the best ones. And in the meantime I'm still just carrying my MM with me and using it to take readings from the TC. Not as elegant, but effective nonetheless. --Ryan

Quick little PS here. Found the converter at Digikey. There are two with identical specs, only difference is package type. What's the difference between 28-SOIC and 14-DIP? What's easier to work with? Thanks, --Ryan

Ok.
So you tried that already, so you will have checked which
of the T/C wires is the braided one, and which is the
signal wire.

Did you make sure that there was only the one track to
ground when you tried it before ?
Namely the thermocouple braid ?

Did you try the braided end (and amp ground) in the other
terminal?

I still feel that the problem is earth current interference.

Yes those readings do seem very low, but as you say, maybe it
only draws current to actually alter the gauge, so you may
only be looking at 'quiescent' current.

I would like you to try one more thing ...

If this does not have the desired result, then i have to
recommend that you use the 'isolator' you speak of.
but i will be disappointed if this has to be done like that.

This may sort it out,
Please let me know how it goes ...

Best of luck with it, John

You'll have to further explain what you mean by braid. You say it's the shielding around the signal wire. Do you mean the metal conduit the wireS run through, or are you distinguishing between the two TC wires themselves. I was under the impression that the yellow wire, NiCr, was the positive lead, and the red wire, NiAl is the (-) and thus would be ground.
Now then, here's what I've done, let me know if it's what you're asking. I've eliminated the ground lead to the the amp and bridged it to each of the following, TC (+), TC (-), and the conduit that the wires run through. All of which had different, very wrong, effects.
One other thing, I read somewhere on a news server forum where someone was having some problem with and amp'd TC also. I can't find the article now so I'm not sure I'm remembering this correctly, or if it even solved whatever his problem was. Someone recommend this guy trying connecting each of the TC leads to chassis ground via a 100 ohm resistor, for a 200 ohm total impedance to ground. Would this have been for a different problem, or similiar. More to the point, what would this accomplish short of lower readings? --Ryan

Hi Ryan,

The amp i am fairly sure, is a balanced input.
By that i mean that it has two input points,
and the output is taken as the difference
between these two voltages.
At least thats what it looks like.

However, i am still struggling to get a clear idea
of the thermocouple wire that comes from the T/C.

It is now starting to seem like three conductors,
a sheath, a red and a yellow.

Could you do a continuity check and tell me the
results, also please describe the wire/cable/lead
that comes from the thermocouple.

Maybe it should be a floating unit, maybe its gone
down on one side?

Also when you use your MultiMeter to read directly
from the T/C which connections give output?
which don't ?

Yes, "the metal conduit that the wires run through"
sounds to me like it, we call that braid or braiding
if its very flexible covering of interwoven thin wires.
Or i suppose it could be a sort of flexible metal ribbed
tubing with the T/C wires inside it.
Please describe it.

This pic is not very good, here it looks like ribbed
tubing.

As for connecting additional resistors on the T/C, no
that makes no sense to me. The output is feeble, 100 or
200 ohms would only reduce it, as you say.

You are correct about the thermocouple wires, but the
amp i think is balanced.
So if you have a leak from one side of the T/C to ground
you can earth that side, and ground that side at the amp.
That may not have come out very well explained.
Please describe the T/C wire.

John

Starting at the washer. SS washer welded to a sort of tube, SS also. The entire length of wire is enclosed in a ribbed conduit. The type that is one long 'wrapped' piece of metal. This conduit is crimp connected to the other side of the tube like connector that's welded to the washer. Running through the conduit are two discrete glass insulated wires, the two TC leads, red and yellow jacketed per industry standard for type 'K' TCs. I can't see where the wires connect, but I'm 99% sure they are bead welded at the washer. That is if the washer were not connected you'd see these two leads bead welded just slightly protruding from the intermediate tube between the washer and conduit. Then the bead is immerssed in the welding puddle that connects the washer to that tubing.
This pic should give you an idea of how the junction looks. Oh, and there's essentially no termination, just stripped leads for the distribution block on the amp. --Ryan
**broken link removed**

Hi Ryan,

I'm now starting to think that maybe that thermocouple was
a balanced type, but has maybe developed a fault to ground.

If the amp has a balanced type input, then the faulty side
of the thermocouple can be connected to the 'sheathing' of
the thermocouple and also to the amp ground, this should
be acceptable to the amp, and remove earth current
interference.

Excellent description BTW.

How did you get on with the continuity tests ?

(three checks)

So,
revised picture on its way, sorry I'm a bit slow .... John

Results from continuity test yesterday (one step ahead). I didn't test the sheathe, rather I went to chassis ground in three different points with identical results. Though these were done with the everything off so readings would be different during operation. Different and inevitably immeasurable by my means since I'm detecting ground currents anyways, they'd throw the MM readings. Anyways, lead to lead on the TC, about 15 ohms. Now then I don't remember which was which, but one lead to ground was reading about 75 ohms, the other was reading about 60 ohms. I'm not really awake enough to do the math, but I'd be willing to bet that those are all correct and the meter is simply being thrown by the current produced by the TC. Kind of an uncertainty principal issue. Act of measuring throws the measurement.
Oh yeah, I had a ball doing quantum physics back in high school. Don't remember any of it, but it was tons of fun. Retired Nuclear Physicist for a teacher. We finished the requirements almost a month ahead of time and had to do something. Even visited the Stanford linear accelerator for a field trip. How's that saying go, know a little bit about everything, and a whole lot about nothing. --Ryan

Hi Ryan,

If you think that the tiny voltage from the thermocouple
is influencing the reading, then just try it both ways round.
The T/C output might be discern-able as a small difference,
but you should be able to judge the mid position.

You sound pretty knowledgeable to me.

Anyway, its the 'sheath-to-T/C' that i am interested in
mainly,
as it is my feeling that you have a fault to ground on one
side.

This fits with the scenario of working OK with an isolated
supply,
but not with its on-board supply in various modes.

So,
If this is the case, then there are two possible faults to
ground.
One wire, or the other wire.

Since i cannot tell them apart, other than by their output,
i suggest trying both situations.

I have drawn an arrow on the 'sheath-and-ground' wire(s)
please try this in the thermocouple terminals, one at a time.

One of them should give the hoped for results.

If this isn't clear, just say ...ok ?

try it on the top terminal, (yellow)
if no good, try it on the second terminal. (red)

What this will do is ground the side of the T/C, if thats the
fault to ground thats ok, if not try the other.
If you're tired then thats ok till tomorrow.
I hope this sorts it out.

I have my fingers crossed.

John

Hi Ryan,

Something else has just occurred to me,
you will probably have to slip some insulation
over the T/C sheathing,
to prevent it from grounding anywhere along its length.

Unless you can just grommet where it passes through,
see what you think anyway.

John

Still nothing. If I reverse the inputs from the TC then the gauge reads negative, actually the needle wraps around backwards since there's no stops. Bridging the ground to either lead still give faulty readings. Rigging the PS is looking more and more like the only viable solution, am I right? --Ryan

Hi Ryan,

Oh yes, you're right, reversing the thermocouple connections
would reverse the output to the gauge.

I'm a bit curious about what you mean by 'bridging the
ground to either lead' for this, it doesn't want any other
ground applied. Doing so would provide another ground loop
and possibly give false readings.

I've drawn the sheathing and the amp (black) terminal joined
and to be treated as one wire, try it on the yellow and the
black, without crossing over the thermocouple wires. By
Using the sheath as ground at the thermocouple end, the fault
becomes irrelevant, so long as it is the only ground applied
to this unit and amp.

You would have to check (visually) that the sheathing doesn't
touch ground anywhere along its length, insulate if necessary,
maybe grommets would do, see what you think.

I have a friend here who is fairly familiar with this air cooled
volkswagen engine, he says its a 'boxer' engine (flat four) and
he would like me to mention some points that you may already
know. He says to change the head bolts for studs, as the bolts
can stretch with the extra load, also to mention that the heat
shield also helps to duct the air, some people remove it, they
think that will help the cooling, but actually it can make it
worse. Looking at the engine from the rear, he says that the
head on the left is the one that generally overheats, and to
pay extra attention to this one. Also that the standard fan
arrangement is not very good, and to consider changing it for
the porsche type fan arrangement, this may involve other mods.
Also he says that a 'Fast-road-cam' is about the maximum lift
reasonably, and that nossing is not an option on those heads,
it would just damage them.

please excuse the writing, i've just re-written this, i had a
TYPICAL 98se crash, just as i was posting, and i lost the well
written version of this. I am still fuming over it. I think
this PC needs more ram, i only have 64mb in there. I hope this
is written well enough to follow, just ask if its not clear!


If this cannot cure your mis-reading problems, then yes, you
will probably have to fit one of those little isolated supplies
that you spoke about before. I'm not very familiar with them.

Best of luck with it, John

Yes that was what I had tried. By bridging I meant hopping the ground from the sheath rather than chassis ground. I've tried both that and a ground wire from the washer at the TC junction itself, still didn't make a difference.
Your friend has a few misconceptions about the engine. First off, all ACVW engines came with head studs, none ever had bolts. Since the case (block) is a soft AL/Mg composite, using bolts would rip the threads out of the case after a few R&Rs. The studs are actually supposed to stretch. When the engine comes to operating temp it will expand from heat, if the studs didn't stretch you'd pound the seating/sealing surfaces for the cylinders at the heads and case. After a few cycles, they would lose their torque and be loose. This is why it's not recommended to go with chromoly studs untill you start boosting (turbo) beyond about 23-25 lbs. Stock studs will hold up to that point, they'll also hold to beyond 12:1 compression.
The tin part is considered common knowledge nowadays, though it wasn't always. I have every piece of tin in place as does anyone who has to worry about engine temps from larger engines.
The 3/4 head (left facing engine) will always run a little hotter due to the air flow vane configuration of the fanshroud, but not much, maybe 20F, #3 cylinder in particular. This is where the TC is located, always measure the hottest part.
As for the fan, it's much better than most believe, provided all oringal tin is in place, thermostat and air flow control flaps are present and working. A Porsche style fanshroud is nice, but leaves few good options for carb linkage, eliminates the stock oil cooler (no problem I run a thermostatically controlled 'extra' cooler), and few are made that fit correctly due differences in engine width from stroking, longer rods, flycut heads, etc. Oh and the biggest problem is they tend to overcool most streetable engines, high CR engines would be an exception. And worst of all they don't accomodate any sort of thermostat for airflow controll. It's always 'on' so to speak.
Funny that he says the heads can't handle nitrous, NOS by the way is a brand name. There are a number of VW engines I know of running large amounts of nitrous with no problems, mind you it's not a cheap setup. All the horror stories about people running nitrous are from people who didn't know what they were doing or skimped on the setup. Here's a tidbit for you. Last year, a flat four VW style engine, Al heads and all, set the world record for quickest and fastest 4 cylinder powered dragster and was the first 4 cylinder engine into the 6's on the 1/4 mile. Before the imports, before Lingenfelter, before them all. All Al engine, well except for the cylinders, crank, rods, and such, but case, pistons, heads. This was pushing at least 65 lbs of boost, exact number is a secret.
This should make your friend laugh and call me crazy. Here's the specs on the cam I run. 260 degrees @ .050" lift, valve lift - .550", did I mention I commute 70 miles/day? If you know what you're doing, anything is possible. I never said it was cheap though, my engine is worth 6x as much as the car itself.

Allright way off topic there. Thanks for all the help John. I did just go ahead and order the dc/dc converter a few minutes ago from Digikey, ran me about $12. I'll have to find some nice way to package it up and setup a voltage regulator for the input, but I think it'll work. However I am still open to ideas because I have some friends who'd like to try this gauge but don't want it to be so involved. The most common gauge to run is a VDO gauge where the TC drives the gauge directly. It's a totally isolated circuit. I just don't like it. Anyways, thanks again. --Ryan

One last thing, take a look at the videos at these links and show them to your friend if you want.

**broken link removed**

The one in the middle. This one is dated about a year. There are two street legal VWs in the 9s now that I know of, brothers actually. Not only that, but they are the guys that built the two dragsters on the next two links, the Lauffer Family of VW Paradise.

http://www.oceanstreetvideo.com/download5.html

The one on top, Kris Lauffer breaking into the 6's.

http://www.geocities.com/crgrapevine/videos.html

Check out Sacramento BOR (Bug-O-Rama) 8-30-03. This is the newest dragster by the Lauffers that shattered the record again just last month. I've seen these guys run. It sounds like a friggin' jet engine going down the track.

Even for those not into the VW scene, these are impressive simply because they are infact VWs. --Ryan

Cheers Ryan,

Some great pics !

John

Woo hoo, I got it working. Well at least it's working on my breadboard, but I did hook it up in the car with the car running and all looked well. And to think I was pissed that after spending all that money on my breadboard I only used it a few times in classe. Guess this made it worth while as it took a few trial and error steps to get it going. I used ended up with a dc/dc unregulated isolating converter at the heart. Finally got a current reading off the amp (helps to set it to dc and not ac, doh!!), 67mA. So my 167mA limit (converter max) is just fine. I did have a little issue when I first hooked it up, I finally got a reading, but the needle was all over the place. Tracked it to two things, both were possible issues mentioned in the spec sheet for the IC. First the output ripple was causing some problems, throw a 1uF cap on the output and that went away. But now I was only getting 6vdc out, when I should be getting a pretty solid 12. There was some mention of having to use a 2.2uF cap on the input as well when for some reason I already forgot. Did that and I have a perfect garbage in, garbage out (1:1) ratio. Gonna have to pick a regulator for the input though cause it's spec'd for 10.8-13.2 input, and I'll be kicking up to 14.4. Luckily I'm still under the "absolute" max input of 15V so I should be good for the mean time. I'll take care of it this weekend. Thanks for all the help guys (John). --Ryan

Hey Ryan,

Good going man !!
glad you're geting some positive results after all the
messing about !

Hope it transfers from the breadboard to the car with
no problems.

Best of luck with it, John

Just did it and it works great. I almost laughed when I thought I finally used something I learned (am learning) in school, but in a field that actually interests me. I just packaged it up on a PC board and project box I got from radioshack. --Ryan

I expect you're pretty pleased about that !
Could you give me the make and model number
of the 12v to 12v isolating unit that you used,
just in case i ever need one!

John