[EE] Need help to ID a part please

[Roger Weichert]

Hi guys,

I am looking for some help please in identifying a couple devices used
across an RS422 communications line on a board.

Ther's a photo here ... http://imagebin.ca/view/UJfdCKe.html

One device goes between one of the RS422 comms line and ground and the
second one goes between the other comms line and ground.

They look just like resistors, but obviously are not. They have what looks
like a glass body with painted on rings. A bit like a large 1N4148.

Out of circuit one measures 194 ohm both ways and the other about 35 Meg
ohms both ways. (no fingers)

I'm assuming they are varistors or transient voltage protectors of some
sort, but have no idea on values or where to source a suitable replacement

The board got damp and had a fair bit of corrosion around that area.

Just out of interest, the 2R2 smd's that are visible in the photo are
inductors, not resistors. You'd think manufacturers could come up with some
markings that weren't so potentially confusing.

Thanks, Roger

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[ivp]

> They look just like resistors, but obviously are not. They have
> what looks like a glass body with painted on rings. A bit like a
> large 1N4148

Might they be TVS or ESD diodes for noise suppression ?

> Just out of interest, the 2R2 smd's that are visible in the photo are
> inductors, not resistors. You'd think manufacturers could come up
> with some markings that weren't so potentially confusing

Yes, I've got some 3R3 inductors in small SMPS, but they are round,
blue and are obviously inductors. Unlike the 3R6 and 1R0 resistors
right next to them. R is an odd choice. Also, inductor marked 101 is
100uH. But so, I've found, might the inductor marked 100. Or might
be only 10uH, depending on who made it

[Oli Glaser]

--------------------------------------------------
From: "ivp" <joeco...@clear.net.nz>
Sent: Thursday, June 10, 2010 12:13 AM
To: "Microcontroller discussion list - Public." <pic...@mit.edu>
Subject: Re: [EE] Need help to ID a part please

>> They look just like resistors, but obviously are not. They have
>> what looks like a glass body with painted on rings. A bit like a
>> large 1N4148
>
> Might they be TVS or ESD diodes for noise suppression ?

TVS or similar sounds like a reasonable bet. Other less likely ideas: they
could be axial capacitors(2300pF?), or maybe a 1N233 diode?
Looks like they are both the same, and the fact you measured different
resistances may mean either one or both is dead or cannot be measured
properly with a multimeter (might be worth applying an increasing voltage
while monitoring current to see how the part reacts). Also, it might be
worth cutting one open to look for a clue.

[jim]

They could be capacitors too.

[Moreira, Luis A]

Hi All,
I need to get a resistor to replace a damaged one on a board where is
used to measure current, it is a 50 mOhm 5W resistor. This is a one off
job.
I found this one http://www.farnell.com/datasheets/6367.pdf . Now this
is a resistor that normally needs a heatsink but I do not have the space
to fit one hence I would like to use it without it. Looking at the
graphs given for derating and calculating for 5W dissipation on it, and
40 degrees C ambient, it seems that I am way bellow the level it would
cause any damage. The problem is that on the datasheet it gives you a
value of 2.5W for power rating on free air that seems to contradict the
graph on page 2.
Am I missing something? Is it because the "flange"(most likely tab)
surface area is too small to effectively be able to dissipate the heat
into ambient?
Any help welcome.
Best Regards
Luis

[Olin Lathrop]

Roger Weichert wrote:
> One device goes between one of the RS422 comms line and ground and the
> second one goes between the other comms line and ground.
>
> They look just like resistors, but obviously are not. They have what
> looks like a glass body with painted on rings. A bit like a large
> 1N4148.
>
> Out of circuit one measures 194 ohm both ways and the other about 35
> Meg ohms both ways. (no fingers)

Assuming you measured this with a normal ohmmeter that puts a volt or less
accross the device, 35Mohm sounds plausible and 194ohm is most likely fried.
I guess they are intended to clamp the lines within some acceptable voltage
of ground. They are probably TVSs. You should be able to replace them with
any TVS with a reasonable rating given whatever common mode range your
circuit has to handle.

********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014. Gold level PIC consultants since 2000.

[RussellMc]

I need to get a resistor to replace a damaged one on a board where is
used to measure current, it is a 50 mOhm 5W resistor. This is a one off
job.

_

Graphs are for flange temperature and Temperature rise Tjc - junction to
case.
You care about Tja - junction to air.

Data sheet implies max junction temperature is 155 C.
At 25C ambient that's a delta T of 120 C.
TO220 pkg has a nominal Rjcthermal of about 60 C/w.
So at 2W you'd use up that 120 C.
They say 2.5 W so that suggests a slightly better Tjcth of 120/2.5 = 48 C/W.

You can improve that quite a lot with vertical mounting and a very modest
heatsink. Chances are you can fit SOME heatsink here - even if not a
traditional one. Thought followed by common sense design should make quite a
lot of difference.

At 40 C you can have (155-40)/48 =~ 2.4 Watts.

Russell

On 10 June 2010 19:28, Moreira, Luis A <Luis.M...@ccfe.ac.uk> wrote:

>
> Hi All,
> I need to get a resistor to replace a damaged one on a board where is
> used to measure current, it is a 50 mOhm 5W resistor. This is a one off
> job.
> I found this one http://www.farnell.com/datasheets/6367.pdf . Now this
> is a resistor that normally needs a heatsink but I do not have the space
> to fit one hence I would like to use it without it. Looking at the
> graphs given for derating and calculating for 5W dissipation on it, and
> 40 degrees C ambient,

[Roger Weichert]

Olin Lathrop wrote

> Assuming you measured this with a normal ohmmeter that puts a volt or less
> accross the device, 35Mohm sounds plausible and 194ohm is most likely
> fried.

Yes, I just used a normal meter. I was convinced that 194ohms was wrong, but
the other one didn't really give me any great clues either.

> I guess they are intended to clamp the lines within some acceptable
> voltage
> of ground.

That was my impression too.

They are probably TVSs. You should be able to replace them with
> any TVS with a reasonable rating given whatever common mode range your
> circuit has to handle.

This is where I am still a little in the dark, because all I have is a
faulty board ... I dont have a circuit or any user manual for any of the
system, or for any of the parameters, other than a legend on the back that
shows it runs on 24v dc. The rest I've had to guess.

Other than not having anything to communicate to the board with, I have
everything else working ... so I'll take a stab at a value, fit a couple
of TVS's, send it back and see what happens.

Thanks to everyone who has replied.

Regards, Roger

[Moreira, Luis A]

Hi Russell,
Thank you very much, that was exactly what I was missing.
Best Regards
Luis

-----Original Message-----
From: piclist...@mit.edu [mailto:piclist...@mit.edu] On Behalf
Of RussellMc
Sent: 10 June 2010 12:33
To: Microcontroller discussion list - Public.

[Olin Lathrop]

Roger Weichert wrote:
> Other than not having anything to communicate to the board with, I
> have everything else working ... so I'll take a stab at a value,
> fit a couple of TVS's, send it back and see what happens.

Take the TVS that measured open with the ohmmeter and put a 1Kohm resistor
in series with it. Now connect a variable power supply to the combination
and monitor the voltage accross the TVS. At some point the TVS voltage will
stay roughly flat as you increase the power supply voltage. This should
happen the same in both directions, but it's also possible for a TVS to be
half fried and only clamp in one direction. Try both directions therefore.
If you get close to the same answer each way, then that's the voltage
rating. You should be able to replace the busted one with a similar or
slightly bigger size part at the same voltage rating.

Power dissipation is mostly a function of the package, so if the replacement
has the same voltage rating and the same package, then it should be roughly
equivalent. Since one of the existing TVSs blew out, it wouldn't hurt to
replace both with a slightly bigger part, since that should take a little
more energy before frying.

********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014. Gold level PIC consultants since 2000.

[Roger Weichert]

Olin Lathrop wrote

> Take the TVS that measured open with the ohmmeter and put a 1Kohm resistor
> in series with it. Now connect a variable power supply to the combination
> and monitor the voltage accross the TVS. At some point the TVS voltage
> will
> stay roughly flat as you increase the power supply voltage.

Thanks for that suggestion Olin, I went straight out to the w/shop to try it
out.

As frustrating as it is, the device appears open circuit both ways. I fitted
a 1K ohm resistor in series and ran the supply slowly up to 50 volts, and
the voltage across the device went straight up to 50V too!! Tried it the
other way too and same thing.

I hooked up the other one, and that acted just like a 195 ohm resistor, both
ways, so I'm pretty definite that one is dud.

> Power dissipation is mostly a function of the package, so if the
> replacement
> has the same voltage rating and the same package, then it should be
> roughly
> equivalent. Since one of the existing TVSs blew out, it wouldn't hurt to
> replace both with a slightly bigger part, since that should take a little
> more energy before frying.

That sounds pretty much how I normally work when I cant find a circuit or
any documentation to help. I normally get away with it, but thought it was
definitely worth a try asking all you guys.

Thanks again, Roger

[cdb]

:: need to get a resistor to replace a damaged one on a board where is

:::: used to measure current, it is a 50 mOhm 5W resistor.

Have you thought of using a short length of resistance wire - can be
purchased (here locally) at 1.73 ohms per metre.

Colin
--
cdb, co...@btech-online.co.uk on 13/06/2010

Web presence: www.btech-online.co.uk

Hosted by: www.1and1.co.uk/?k_id=7988359

[Michael Watterson]

cdb wrote:
> :: need to get a resistor to replace a damaged one on a board where is
> :::: used to measure current, it is a 50 mOhm 5W resistor.
>
> Have you thought of using a short length of resistance wire - can be
> purchased (here locally) at 1.73 ohms per metre.
>
>

At 5W it may light up with a cherry glow before the solder melts and it
falls off, or sets something on fire. Also Temp Coef. may not be good
for measuring current.

[ivp]

>> Have you thought of using a short length of resistance wire - can
>> be purchased (here locally) at 1.73 ohms per metre.
>>
>>
> At 5W it may light up with a cherry glow before the solder melts and
> it falls off, or sets something on fire. Also Temp Coef. may not be
> good for measuring current

Several strands in parallel would get down to 0.05 ohms and add bulk.
I've done this a couple of times to make medium-power shunts

For example, 13.77 ohm/m nichrome (7.26cm/ohm). 0.05R is only
3.63mm but 10 strands 3.63cm in parallel -> 50mR. Or 20 strands
of 7.26cm

Nichrome or Cuprothal will have published temperature coefficients
that can be used to assess whether changes are significant

[RussellMc]

You can get Nichrome in many gauges. Can also parallel strands.
I have used it at 500 Watts and no visible temperature effects (maybe
50C? by suitable choice of length and diameter.

Also can get Constantan which has higher resistance per square (per
length at given area) and is almost invariable in resistance with
temperature. Has a commoner name which flees but will return and
Gargoyle knows. Also cheap.

Can rate wattage for whatever temperature rise suits.

Neither solders nicely with standard solders.

R

[Michael Watterson]

RussellMc wrote:
> You can get Nichrome in many gauges. Can also parallel strands.
> I have used it at 500 Watts and no visible temperature effects (maybe
> 50C? by suitable choice of length and diameter.
>
> Also can get Constantan which has higher resistance per square (per
> length at given area) and is almost invariable in resistance with
> temperature. Has a commoner name which flees but will return and
> Gargoyle knows. Also cheap.
>

I've not heard another name for it.
However Google suggests maybe a *Less* common name is Eureka
http://www.resistancewires.co.uk/products_RW45_Eureka.html
Nickel 45% Copper 55% (same alloy)

> Can rate wattage for whatever temperature rise suits.
>
> Neither solders nicely with standard solders.
>
>

I've seen it riveted.
I wonder what kind of solder works?

Obviously a fatter, longer piece is less likely to act as "filament" or
Polystyrene cutter. Teflon or fibreglass can be used to insulate I suppose.

http://en.wikipedia.org/wiki/Constantan
This may *now* mention Eureka (one never knows if someone higher in the
food chain takes exception to an edit)

[ivp]

> Neither solders nicely with standard solders.

Can't say I've noticed that with the resistance wire from DSE
(Cuprothal 6R/m) or Electus (Nichrome 13R77/m)

[Olin Lathrop]

Michael Watterson wrote:
>>> need to get a resistor to replace a damaged one on a board where is
>>> used to measure current, it is a 50 mOhm 5W resistor.
>>
>> Have you thought of using a short length of resistance wire - can be
>> purchased (here locally) at 1.73 ohms per metre.
>
> At 5W it may light up with a cherry glow before the solder melts and
> it falls off, or sets something on fire.

Do the math. At 1.73 ohms/m, you would need 29mm (1.1 inches) to achieve
50mOhms. At 5W that would definitely get hot. You wouldn't want to touch
it. But melting the solder at the ends is a bit melodramatic. A good
fraction of the heat would conduct straight from the wire into the air.
Even if none of it did, each solder joint would have to dissipate only 2.5W,
so 2W or less is a reasonable figure. Again, that's going to get hot, but
well below solder melting temperature. Picture it this way, how far do you
think you'd get trying to solder the wire on with a 2W soldering iron?

You could use two sections of twice the lenth in parallel, but spaced apart
so that they are somewhat separately air cooled. Now you've got a total of
about 4.6 inches of wire and 4 solder joints to dissipate 5W. Again it will
get noticeably warm, even "hot", but not beyond anything the wire, solder,
and a ordinary circuit board should be able to handle.

********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
(978) 742-9014. Gold level PIC consultants since 2000.

[Michael Watterson]

Olin Lathrop wrote:
>
>
> Do the math. At 1.73 ohms/m, you would need 29mm (1.1 inches) to achieve
> 50mOhms. At 5W that would definitely get hot. You wouldn't want to touch
> it. But melting the solder at the ends is a bit melodramatic.

In this case, I agree :-)

> A good
> fraction of the heat would conduct straight from the wire into the air.
> Even if none of it did, each solder joint would have to dissipate only 2.5W,
> so 2W or less is a reasonable figure. Again, that's going to get hot, but
> well below solder melting temperature. Picture it this way, how far do you
> think you'd get trying to solder the wire on with a 2W soldering iron?
>

No, I didn't do the maths.
(Though how much heat of an 11W iron goes into body of iron and air and
how much on the joint?)

> You could use two sections of twice the lenth in parallel, but spaced apart
> so that they are somewhat separately air cooled. Now you've got a total of
> about 4.6 inches of wire and 4 solder joints to dissipate 5W.

There may still only be two joints. but assuming the wire can be
soldered, it's probably OK.
I think I may have seen crimped on pins on non-copper wire on thermal
binders etc.

> Again it will
> get noticeably warm, even "hot", but not beyond anything the wire, solder,
> and a ordinary circuit board should be able to handle.
>
>

I have evil childhood memories of the Dropper Resistors of old All
Valve (tube) TVs. I think on some they used higher MP solder as on at
least one occasion the solder dripped off on a replaced dropper.