Substitute for UA78MGU1C

[gilbenl]

Anyone know a substitute for the UA78MGU1C or LM78MG and its negative counterpart? 4 terminal adjustable (5-30/-30--2.2) voltage regulators

http://www.alldatasheet.com/datasheet-pdf/pdf/161847/NSC/LM78MG.html

Thanks in advance

[mariush]

They're plain linear regulators like a 7805, but adjustable instead of fixed.

LM317 and LM337 should be suitable, though they're only three pin ( input, output, adjust)  while your LM78g have ground, input , output, adjust ... you just leave the hole for ground empty.

[dacman]

According to nteinc.com, it crosses to an NTE953, which is available at both alliedelec.com and onlinecomponents.com.

The NTE954 was the counterpart, but it apparently isn't made anymore.  NTE presently makes the NTE1927 in a TO-3 package.

[SeanB]

Still have one or two of those on a board somewhere, came in an old coin counter I got a long time ago on auction ( not working).

[gilbenl]

Thanks for the heads up on the NTE953. Also on eBay for $10.50 shipped as of this post. Not on eBay, but google came up with the negative counterpart-NTE954

http://www.electronicsurplus.com/nte-electronics-nte954-vr-neg-adj-2-2-30v-1amp

Thanks for the help folks

[macboy]

They're plain linear regulators like a 7805, but adjustable instead of fixed.

LM317 and LM337 should be suitable, though they're only three pin ( input, output, adjust)  while your LM78g have ground, input , output, adjust ... you just leave the hole for ground empty.

The built in voltage reference of the LM78MGC is 5.0 V, compared to 1.25 V for the LM317. This affects the R1/R2 adjustment resistor ratio.

If you want to keep the resistors the same, you can substitute the LM7805, with it's ground pin used in place of the LM78MG adjust pin. The main difference here is that the quiescent current of the LM7805 (5 mA or so depending on manufacturer, type, etc.) will flow through the lower set resistor (R2). This will cause some error in the set voltage. The lower the value of R2 the better, as the error voltage is directly proportional to it. The 78L05 will be better as the quiescent current is lower, but consider the required current and power dissipation. The quiescent current may change with input voltage, load and/or temperature so consider the resulting errors. As long as this error is acceptable, then it is probably a reasonable substitute. My guess is that the resistors will be high enough in value to cause too high of a voltage error. Using LM317 is better since it has far less current through the Adj pin (its quiescent current flows to the output pin instead) but you would definitely need to change R1 and/or R2.

[gilbenl]

This might be helpful...

Seems like the 7805 would be a second choice to the NTE953. I'd prefer to not to mess with the existing resistor setup.

[macboy]

I don't think that the 7805 will work well here. The resistors are too high in value, and the voltage drop across them due to 7805 quiescent current will cause significant problems.

If you just need +/- 15 V, why not use 7815 and 7915? The resistors and adjustment pot would no longer be used. Be careful of pinout differences obviously.

[dacman]

I recognize that circuit.  I've replaced W2 a few times and I've replaced a few trimmers (although neither R7 nor R12), but I've never seen those regulators go bad.  May I ask what the issue is?  Voltage levels may seem to point to a bad regulator when the problem is actually W2.

[gilbenl]

@dacman-I bet you do recognize the schematic...Would've asked you directly, but didn't want to lean. This is related to the second 1062 with the unstable reading.

Following your suggestions, I found significant ripple on the +15 compared to the -15 rail. Could that be W2 or am I correct in assuming its the LM78? If you check the other thread, you can also see what other steps forward, and steps back, I've taken since my last post.

https://www.eevblog.com/forum/testgear/datron-1062/msg680397/#msg680397

[dacman]

About the crosses, NTE adopted the same part number convention as ECG, and both NTE953 and ECG953 are the same and both NTE954 and ECG954 are the same.

The rectifier is a hard working part, because its diodes are continually turning on and off.  The diodes can get weak and they can blow.  W2's generic part number is W01G, but a W02G or W04G would work fine (the number can go up, but not down).

How much greater is the ripple on the +15V rail?  1062s can be temperamental.

[c4757p]

The rectifier is a hard working part, because its diodes are continually turning on and off.

erm.

[SeanB]

Rectifier diodes cook because they are generally running with a high peak current as the filter capacitor charges, limited only by the resistance of the transformer and the ESR of the capacitor. the average current ( over a mains cycle) is low, within the rating of the diode, but the high peak currents do degrade the diode with time. You often find that the diodes run very hot, even though the current seems to be low. The heat pulses cycle the diodes and eventually they degrade and fail.

[Ericho]

I've rebuild a few old PS boards containing the old 78MGxxx/79MGxxx

I Always used 78xx/79xx fixed regulators to do the job.

It involves crossing some of the TO220 leads (pinout), removing a few resistors and a few jumper wires.

Must be 70' or early 80' gear you're fixing so replace the Elco's as well while your at it 

Kind regards,
Eric

[gilbenl]

Might be a trivial point, but say I use the 78xx/79xx, which would require crossing some leads. In this particular application, the mounting distance (board to heatsink screw) is rather short and there is limited distance between this reg/heatsink and it's neighbor. Is there a "preferred" method of mounting/crossing? My first thought would be to mount it with the leads away from the board and use some short bond wires to connect it.

That being said, once I have the filter caps and the 309K replaced I'm going to retest at the rectifiers before I go down this rabbit hole. Going with any of the few variable regs is going to cost me $10+sh per regulator.

[dacman]

There are two voltage clamps on the analogue board in the 1062, D73 and D74.  Their VR is 15V.  Fixed regulators may not regulate at exactly 15V.