Cheap Multimeter: is this thing giving me correct hFE readings?!?

[Earthscum]

So, I got a cheap Cen-Tech 7 Function Multimeter to get a round-about reading for transistor gain. It is here: http://www.harborfreight.com/cpi/ctaf/Displayitem.taf?itemnumber=98025

Ok, so I get it and plug in some 2N's. 2222's are giving readings around 27 or so, while 4401's are giving up about 22, and 3904's are hitting 5. WTF? I tested Ohm to Ohm against my Craftsman. On a 1M pot maxxed out, I was reading 1.047M on the Crapsman, and 1.07xM on the Cen. When I did a smaller reading off a 1k, the Crapsman is reading 12.35R and the Cen is reading 13.34!

Wish I could afford decent equipment like a Fluke, but unfortunately I can't.

So, all in all, I just want to get a decent reading out of these things, and was wondering if this is in any way accurate, should I open up the case and try calibrating it? I'm sure I can get my hands on a GOOD meter, being that I live in a college town (CSU, Ft Collins).

I guess I'm just wondering what's up with this thing... I can kinda understand the 3904's being lower gain if this is doing the test with ultra low currents... the 4401's would show more gain at a lower current than the 3904. I even pulled the trannies out of a working pedal and tested. The pedal works just fine. Also tested a TIP120... lol, it read "1".

Maybe I'm just missing something? BTW, R.G.'s tester doesn't make sense to me... of you switch around the battery and tester leads to test NPN's, then you are running your current backwards?!? I've been searching for about a total of 5 hours now trying to test gain, going through a bazillion resistors, testing, etc. This is frustrating me. Also out a couple transistors from the manual testing, which is why I bought this crap-tech meter. Basically, extremely limited on space and budget. $10 max type of thing, and GF is constantly pissed about the pile of crap on the living room floor, so I really don't need another dozen testing modules laying around... already have a couple for testing JFets and Mosfets.

Again, I already know 90% of the responses are gonna be "quit buying crap" - "buy a Fluke" I know. If I could afford it, I would be in a $500k house, paid for, with a room full of O-Scopes, comps, sig generators, etc. but the reality is that I can't afford a $100 pedal, so I decided to learn how to build my own.

So, any idea what is up with this thing? do I need to take into account the current it's using, or you think it's just that much of a turd that they put in a basically fake module for selling point?

[Ripthorn]

I have the same meter for testing hFe and it works fine for me.  My BC108's were measuring about 550, which is in the specified range for them (C's).  I would think either you've got leads swapped around or some other simple thing.

[Earthscum]

hmmm... well, my buddy has the same one and he's gonna bring it by so I can test side by side.

I'm kinda thinking that I got  a flop meter, considering the readings I was getting. The diode tester  between the 2 meters I have is only differenced by  <.005V. I also tried out my Mac DVM  against the Crapsman, and they are within ~.1% of each other. I just hate my Mac because it has a bad LCD, and I can never track down the Mac guy to get it replaced, even though there's an Aamco next door. Lol, I have the same problem with my Matco tools, though... hard to track the guys down when you aren't in the business anymore.

[George Giblet]

When you see readings of 5 and 20 it usually means you have the pin outs wrong, ie. C and E reversed.

There's nothing magic about transistor testers, there's nothing in them, a couple of resistors like RG's circuit.   Regarding the PNP/NPN you can do the supply meter/terminal swapping with a DPDT switch, it's very simple.   The modern meters don't use a switch.  In order to remove the switch they use a different test configuration which is slightly inferior because it runs the two types of transistors under different conditions.  I'm fairly certain the schematic for that meter is here,  the transistor tester is in the middle the PNP test is the same as RG's but the NPN is different.  IIRC the Vcc supply for the tester section is about 2.5V,

http://www.hobby-hour.com/electronics/m830b-schematic-diagram.gif

You should place too much emphasis on accuracy of DMM's diode or transistor gain test ranges.  In both cases these aren't single numbers the number vary depending on the test conditions and the test conditions vary from meter to meter - although the test current is typically agree within a factor of 3.

Some transistors have a flat gain curve,  eg. fig 2 of,

http://www.nxp.com/acrobat_download/datasheets/BC549_550_4.pdf

Where as others have a widely varying gain curve, eg. fig 13 of,

http://legwww.epfl.ch/~decurnex/TP_Electronique_Master_2/Data%20Sheets%202/2N2369A.pdf

The curve shape will vary from transistor to transistor.  So if you match the gain of two transistors using the transistor tester (at the tester collector current), your circuit might operate them at a different current and the gains could then be different.

The diode voltage on the other hand depends on the current through the diode, the diode exponential law.  If your test current varies 3 to 1 from meter to meter then you could see meter to meter differences upto 50mV *on the same diode*.

As far as buying crap goes.  I've found those meters do produce good readings (AC ranges can have some variation) I have compared a number of these to flukes etc.    I've read some meters aren't they great for accuracy, but I haven't come across any, and comanies offere re-calibration services.  I have noticed some produce dodgy measurements when the battery is flat, even though the flat indication hasn't kicked in yet.   I actually have one similar to the one you posted.  The main problem areas are the case breaking when you change batteries,  the switch going dodgy, and the meter possible failing.  Some use very cheap leads and the type is slightly different to quality meters.  If I was producing electronic equipment I suspect the switches would fail after a while.  Of about 6 cheapish meters I've only seen one actually fail and that was because of a bad solder joint on the PCB - which I managed to fix with some effort.  One the case broke.

Recently we bought this one as a "kick around" for work,

http://www.jaycar.co.nz/productResults.asp?FORM=KEYWORD

The case is quite strong and the switches seem very solid  but it's somewhat more expensive. (no gain though)

Personally I wouldn't go buying a decent meter just because it had a gain measurement but buying a cheap arsed one just to do gain measurements isn't a bad option if you are going to use it - not worth buying the switches and box in some cases.

[station]

Also, make sure you have a full battery. Sometimes that will skew the readings.

[tackleberry]

I bought one of those same cheap meters at a harbor freight store. Went to use it and it wouldnt give a hfe reading with any trans I had no matter how the legs were tried. Went back for an exchange a second 1 I tried before leaving the store didnt work at all, spurged and bought the next model up and it seems to work better. The ohms readings between it and my fluke are real close, my fluke just doesnt have any diode, cap or hfe test settings.

[zombiwoof]

Wow, when you said cheap multimeter you weren't kidding!  $2.99!  You'd think the components in there would cost more than that.
Ain't modern technology wonderful?

Al

[puretube]

Recently tried to "Hfe"-match a couple of pairs...
one "DMM" gave me ~280,
the other ~330;

both were right, obviously: ~305 +/- less than 10% ...

[jonfoote]

i've got one of them, you have to wiggle the transistor about in the socket for a bit till you hit the sweet spot.