So I had a battery leak into my Heathkit IM-5228 vtvm. Cleaned it up and replaced the battery but it made me think about a battery eliminator (BE). Implementing such a thing requires some careful planning and design, especially concerning thermal management. Not so much in terms of heating the box up. Nominal power use in the device is 10W, and almost any conceivable BE will dissipate 1W or less. But an LM317 dissipating 1W requires a heat sink. This would only occur when measuring ohms in the Rx1 region, where the 1.5V is applied to the resistor being measured and a 9.1ohm range resistor in the cabinet, and then only when a test is being done. I never have a resistance check that takes more than a few seconds.
I thought a lot about this – way more thought than the application deserves. But in thinking about this and noodling around with designs I learned a lot. In any case, the discussion below is admittedly over the top. But here goes.
Most of the designs are based upon the venerable LM317. Some simple circuits tap the hot lead from the vtvm filament transformer tap (6.3vac). And the BE output is 1.55v. If we are measuring a near-zero-ohm resistor then 1.55v into 9.1 ohms gives a current of 170ma. Dissipation in the range resistor is i^2r, or .26W. Dissipation in the LM317 would be (6.3-1.55) * .17, or 810mW. Assume 1W, so we need a heat sink. Altenately, could we put a 6.8 or 7.5 ohm resistor in series with the power input? With .17ma current, this drops the supply voltage by 1.3v, so that LM317 dissipation is ((6.3-1.3) – 1.55) * .17 = 0.58W, and the resistor dissipates .22W. The idea is that the power dissipation is not in the LM317. This still is borderline as to whether a heatsink is required.
My current design is shown below. Note that in some versions of the IM-5228 manual online there are patently wrong circuits, and other designs where the circuits could operate correctly but are lacking in some elements (e.g. protective diodes). What do you all think?
[Apparently I can’t add graphics files…. Sorry]
Some LM317 questions:
1) Can I put the resistor in series with the LM317 input to offload some of the power out of the LM317?
2) Does the LM317 provide a stable enough, noise-free output?
3) Can I safely tap into the filament tap on the Heathkit transformer or do I need to add a transformer?
4) Some designs use a single diode for half-wave rectification. It seems to me that this puts a higher stress on the transformer filament windings. So would a full bridge be better? It’s only 3 diodes more.
There are a lot of really great resources online. Of course, Mr. Carlson is one of them. Others are:
a) Dave Jones, of the Australian EEVBLOG (highly energetic discourse on diagnositic tools and general topics)
b) Mark Maher, of Mend-it-Mark (Expertly diagnoses and fixes consumer electronics both recent and vintage)
c) Jonn Teel, of Predictable Designs (Chip centric (Teel designed chips for TI) with good advice for product design
d) Hans Rosenberg, more good advice about designs, emphasizing high performance topics (e.g. eighth order filters and resonance loops forming between layers of multi-layer boards).
In one video, Teel points out chips that you shouldn’t use for current designs. One such chip is the LM317. Fine for one-offs, but for products being made on a Grand scale it has liabilities like a large (3V!) drop out voltage. So while Im sure that the LM317 design above (or some tweaked variant) would work, I started looking at other voltage regulator chips, such as TIs 767TLV. In designing a 1.5v supply based upon this I thought that I could take advantage of the lower drop out voltage and possibly use a 115vac to 2.5vac (center-tapped) to lower thermal load on the regulator chip (yes, I realize that a whole transformer is a bit over the top here, but I’m pretty sure it would work really nicely). My candidate design is below. Comments? Criticisms? In addition, some questions.
[Again I can’t add graphics files…. Grrrrrr]
5) Is it worth it to use the more modern 767TLV chip? It looks like about the same cost either way.
6) Again, if I want to limit the dissipation in the chip itself, my design has a transformer. Could I instead just use the filament voltage with a bridge (and a series resistor)?
I’m leaning towards using the 767TLV with filament voltage, a full bridge, and a dropping resistor.
Hi regarding the “inability to add graphics here” yes it is 100% doable but because the site itself it is not meant to store/host any user graphics (IE: pictures, PDF docs txt docs, etc) you’ll need to use some third party image hosting site and the unanimously suggested one for the use in here is IMGUR. Here is someone’s posted instructions but somewhere on this forum MCL itself has posted a little HOW TO post on IMGUR but i have no idea where to find it:
https://mrcarlsonslab.com/community/postid/453/
and if you post the link containing the image in a certain way no clickable link will appear (to re-direct you to the picture hosting site) but instead you would directly see the picture withing your message as if the picture is attached your message. I’m not entirely sure of the type/kind/name of the link but i think is called either embedded link or direct link (not sure which) On the other hand people have used other hosting sites not only IMGUR They even shared pictures and PDF files using their own google drive or even some picture hosting which don’t even require an account to allow you to post pictures and the one that just popped in my mind is postimages.com er…but like I said, there are many others that would or at least should work just fine.
@ovi4 Thanks. I uploaded to google photos, and linked the pics to some text, and the website deleted the reference.
I don’t think that I’m smart enough to use this forum. Sigh.
Check out the link on EEVBLOG.
New members posts need to go through manual review before you’re able to post links. The system will automatically delete them. You’re an active member now so it should work.
-Mr Carlson’s Lab Admin
Thanks, Mr. Carlson. Below are links to the two schematics. The feedback I was looking for is this:
The max current that the BE needs to supply should be about 170ma. If I use the 6.3vac filament secondary 1) Does this put undo stress on the filament secondary? Or do I need to add in another small transformer, like a 115vac to 2.5vac CT design?
The 6.3vac is rms, which is about +/-7.5 volts peak (6.3 * 1.414 – 2 x .7). If I drop this to 1.55vdc using an LM317 or a 767TLV, this develops about 1W in the worst case. This will require a heat sink. Not a big problem, but 2) is there a way to put a resistor in the rectified 6.3vac (e.g. 7.5vdc) supply to the regulator chip to dissipate some of the drop outside the regulator chip? Say an 18 ohm resistor? Does that do anything useful? Checking a 3 ohm resistor is not a frequent measurement, nor does it take a lot of time!.
Would the experts here use a cheap, tried and true, but dated LM317 design with a high drop out voltage requirement? Or the newer 767TLV?
https://photos.google.com/photo/AF1QipMTk-a-qw-W2EDNmJbQY2w7usUgA8_tpRgbWdcS
https://photos.google.com/photo/AF1QipPQvZo_AO6WsuXSF7wXAydkg_NT6C52XlTaohje
So I created links. Let’s try this:
https://photos.app.goo.gl/4oVqWahuKfWRc9uC9
https://photos.app.goo.gl/iVSMPECAmgoDi6X39
Do these work for you guys? Thanks.
The LM317 Battery Eliminator design above is defective. I tied a full wave rectifier to the filament secondary. But both one leg of the transformer secondary and one of the DC outputs of the bridge are grounded – a good way to burn out the transformer. So one must use either a single diode half-wave, or a separate xformer without either secondary leg grounded.
Here’s a working idea (I think…)
https://photos.app.goo.gl/nHLknciFXXYy4guL8
@wizard-of-boz a few comments:
- not sure what D2 is for. D2 will prevent C3 from charging. This will make C3 ineffective
- have you looked at the footprint for the 767TCL? You will need special equipment and skills to install that IC
- With a half wave rectifier, don’t expect the 6.3v to be solid. I would think the half wave dc would droop below the LM317 3V dropout and cause havoc
@frogman Thanks – useful comments. The D2 and D3 diodes are suggested for protection in cases where the inputs or outputs are shorted and polarity switches rapidly on the terminals. See this diyaudio post for the idea. Overkill here.
The point that really made me think was about the thermal management. Much easier to put a heatsink on an SO-220 LM317 package!
But WattsThat on EEVBLOG pointed out that Heathkit had a solution here, and its way simpler and cheaper. At the cost of some variability in voltages when you use different ranges. Check this out:
https://photos.app.goo.gl/2rumxNPkKYUhTqTs6
The original sketch had a 27R 2W resistor. My calcs seem to indicate that a 33R resistor works better, yielding delivered voltages between 1.5V (lowest range, 0R measured) to 1.7V (higher ranges).
In thinking about the Heathkit solution, it is clear than in the low ohms setting, that the value of the tested resistance will affect the voltage delivered. I have to think a bit about the sensitivity of that voltage, and its impact on the ohms measured. Certainly a regulated near-constant voltage would not have that problem. Hmmm.
@wizard-of-boz ok after looking at the recommended DIY site, I see the D2/C3 combination is connected to the adjustment wiper arm. That wasn’t clear in the original schematic. Disregard my comment about that
@frogman, will do. The heat management point you made was helpful and decisive, though. Thanks.
So I implemented the LM317 regulator and installed it. Initially I was going to use pop rivets to attach it to the back of the internal chassis but this created about a 0.5mm interference. Saw a post by Jeff Trantor where he removed the battery cups. Didn’t want to do that but Jeff’s install (of the Heathkit resistor plus two diodes approach) was clean. The set up works. I used 3 dropping diodes, not four.
https://photos.app.goo.gl/sWRDPYWnW7Ayg3tE7