-5V

Muerto

Well-known member
Joined
Mar 6, 2008
Messages
1,615
Reaction score
40
Location
Invisible, smoke, Denmark
I have changed a componet (a 7905C -5V) with a 7905CV -5V, but the thing is now it´s reading -12V...??? - can the game vork with that, or does it has to be -5V (+-1V or 2V??)
 
I also have that happening on an A/R II I'm using for our Centipede. The -5v is at -11v and I've changed the 7905 three times already.

I ordered a bunch of different 7905 regulators from Mouser as well as one from Bob Roberts to see if any of them help.

On Centipede the -5v is just for audio and it goes to the LM324. The original LM324 blew but after I replaced it it's been running fine with -11v. I scoped the audio out and it's offset by -5 to -6v... not sure if that's because of having -11v driving the chip or what?

I was hoping to test the different regulators last week but that didn't happen and we are on vacation this week so it won't happen to at least next week.

If nobody pipes in with what is going on by then, I'll be sure to post here or PM you if/when I figure it out.
 
There are a bunch of bad 7905's out there from what I have seen. I buy the 2% tolerance 7905's from Mouser and have not had a problem yet.
 
Make sure the grounds are all at the same potential. It sounds like the ground on the -5V has floated to -6V giving a differential of -11v and an apparent of -5V.

ken
 
Could this be a Bob Roberts "Chicken Little" issue? I was not happy to read Bob's page about this because I have always depended on the voltages to measure at their expected values even without a load. Now I have to add a load of the right amount (I am no EE, so not trivial for me) or risk that the voltage is really bad and damage PCBs when I plug them in!

http://arcadecontrols.com/BBBB/7905.html
 
Could this be a Bob Roberts "Chicken Little" issue? I was not happy to read Bob's page about this because I have always depended on the voltages to measure at their expected values even without a load. Now I have to add a load of the right amount (I am no EE, so not trivial for me) or risk that the voltage is really bad and damage PCBs when I plug them in!

http://arcadecontrols.com/BBBB/7905.html

Hard to say. I love Bob to death, but don't use his 7905's any longer.
 
Easy to add a load with a clip on light - just use a small 12v lamp - then take your measurement.

I had to add "regulation" (a string of 10 1N4004's) to a switcher to bring its -12v to -5v (10x.7v drop=7v drop) for an OLD 486 motherboard. It showed -12v until I loaded it with a resistor - I think I used a 10K value or some such thing - then the voltage read just fine.
 
7905 should regulate relatively close even without a load. Bob is wrong.

I would have agreed with this....until I got a couple from Bob. I don't know how they do it....but they don't regulate worth a damn without a load. Put a load on them....and they (seem) to be fine. I don't get it.

Edward
 
Yeah you shouldn't need a load. Now on some switchers you will need to load the 5v output to get it to regulate because of the design. But what you are dealing with on an AR board is a simple linear supply, that's regulated.

I would definitely check to make sure you don't have a bad ground, Are you measuring the voltages straight from the 7905? Or test points?
 
The 7905's Bob sells must be garbage. I've never seen a regulator need a load to stay within a couple hundred millivolts.

I would have agreed with this....until I got a couple from Bob. I don't know how they do it....but they don't regulate worth a damn without a load. Put a load on them....and they (seem) to be fine. I don't get it.

Edward
 
I would definitely check to make sure you don't have a bad ground, Are you measuring the voltages straight from the 7905? Or test points?

I measured from both with two DMMs and a scope(two scopes if you count a cheap USB jobber). I didn't look for grounding issues so I know what I'm going to do first when we get home from vacation. If that leads nowhere, then I have 5 different brand 7905s to test.

And I agree with everyone regarding the loading requirement... on a regulator it just shouldn't be necessary. :(
 
The 7905's Bob sells must be garbage. I've never seen a regulator need a load to stay within a couple hundred millivolts.

OK - this has my curiosity peaked.
Does anybody know the brand of these bad 7905's?
Any manufacturer markings on them? ST?

Ed
 
I still have one of Bob's, and it is marked as follows:

f1 J29
LM7905C

Guessing from what you have - it sounds like a Fairchild LM7905CT which is one of the better ones you can get.... as would be the National Semi version with same number.

But, Bob's website says LM7905CV... no clue on who this would be. Amost sounds like an ST Microelectronics number... but their number would be L7905CV without the "M". The ST parts are cheaper and don't quite have as good of tolerance but still should be well within spec as long as minimum loading is present.

Looking at datasheets - all of these linear regulators DO have a 5mA minimum load current. If there is no load on the output - the output is unregulated and often undefined. Many DMM's can read a real high reading as they present nearly no load to the power supply. Some analog meters are more of a load and will read closer to normal value.
These regulators need about 5mA to attain proper regulation. For the most part, a simple 10mA LED indicator would have done the trick. If you are testing a power supply with no load then it would be advised to put a small load resistor on the output. A 470ohm resistor *should* do the trick.

Ed
 
7905 should regulate relatively close even without a load. Bob is wrong.

Check the data sheets. 7905's as well as many other linears DO need a minimum load. Typically about 5mA so it is minimal. But if you have a supply that just outputs from the board without even an indicator LED then it is a poor design and not designed to the manufacturer's specs. As long as the -5 regulator has a load such as an IC - then the regulator will be pulled into spec.

These regulators have internal comparators for controlling the output voltage. The comparator input voltage is based on a pair of voltage divider resistors. If there is no output current flowing (no load) then the comparator/regulator circuit cannot function and the output voltage will be undefined.

Ed
 
Looking at datasheets - all of these linear regulators DO have a 5mA minimum load current. If there is no load on the output - the output is unregulated and often undefined. Many DMM's can read a real high reading as they present nearly no load to the power supply. Some analog meters are more of a load and will read closer to normal value.
These regulators need about 5mA to attain proper regulation. For the most part, a simple 10mA LED indicator would have done the trick. If you are testing a power supply with no load then it would be advised to put a small load resistor on the output. A 470ohm resistor *should* do the trick.

Ed

I'm in agreement with this all the way:
http://forums.arcade-museum.com/showthread.php?p=841763#post841763

The output voltage is only specified if the output current is between 5mA and 1A.
There may be some regulators out there that will show a solid -5V with almost NO load (or a small internal leakage).

I see the problem though. Without this knowledge, at first glance it looks like there is a problem when the output voltage measures so far from what it is expected to be.
 
Yet another reason why I trust my "old school" Simpson 260 over any of my DMM's.
:)
It would seem that the 260 would put enough of a load on the regulator to get a proper reading no?
 
Yet another reason why I trust my "old school" Simpson 260 over any of my DMM's.
:)
It would seem that the 260 would put enough of a load on the regulator to get a proper reading no?

Holy Christ! I haven't even seen a Simpson 260 in 20 years, seriously. We used them back in the old days, before we went to the Fluke 77 A/N
 
Back
Top Bottom