EEVblog 1586 - Stranger Things Decade Resistor Box

[EEVblog]

A $12 6 decade resistor box, is it any good?
https://s.click.aliexpress.com/e/_Ddfe4PL
https://s.click.aliexpress.com/e/_DnK2irb

[Kleinstein]

The switches may not last very long, but for less frequent use it may be OK. In principle the switch design looks good.

It is a bit strange they actually used 5 ohm / 50 ohm and so on resistors and not 2 x 10 ohm, 2 x 100 ohm and so on in parallel. This would have reduced the number of different resistors and given the 5 resistors at least a little more power handling capability. The 5 ohm is also actually 4.99 - so starting with a small error already, though likely less than the nominal tolerances.

At least it looks repairable, if one burns a resistor.

The .1 ohm are indeed of limited use, though not totally useless. It would be resonable easy to just change that decade to 1 M steps.

[EEVblog]

The switches may not last very long, but for less frequent use it may be OK. In principle the switch design looks good.

Most decade boxes only get occasional use.

[/quote]The .1 ohm are indeed of limited use, though not totally useless. It would be resonable easy to just change that decade to 1 M steps.[/quote]
Yes, I think I'll mod mine to 100k steps.

[Macbeth]

I wouldn't be so sure. I have a 1970's ESI Dekavider that was clearly used by the Military Industrial Complex contractors by the calibration stickers and also a label stating that it should always be set to a particular setting for its job. Essentially a very expensive potentiometer in its day.
The switches are fantastic, double sided wipers, solid brass on gold plated contacts. The resistors are all wire wound and can cope with much power and high voltage.
This Chinese box looks just like a traditional Bakelite resistor box which would be of similar construction but is totally fake.
I have built a few tiny resistor boxes using digital decade switches and a 3D printer for the case. They are great but 0.1 ohm is ludicrous. Also I used 1W resistors on the 1 ohm range just in case. Lots of examples on EEVBlog forums.
Having said that for people who still deal with ancient valves/tubes then this is a great box to stuff some high wattage resistors in. The useless binding posts have to be replaced.

[PartialDischarge]

I had one of these bought some time ago, and to make it short: To anyone owning this decade box, put it in the trash, it’s the best advice I can give.

[rs20]

One thing that has always bothered me about these decade boxes is that they have a substantial offset; for example 0.4 which Dave zeroes out at the start.

But that means that if you want, say, 100.0 , you need to type in 99.6 . I can't help but think that there must be some clever way to incorporate this right into the mechanics of the box, such that if you ask for 0.0  , you get 0.4 , but if you ask for 100.0  , you get 100.0  . But I'm not clever enough to figure out such a trick. Is there a way? Other than using a microcontroller and 20 relays in a binary arrangement

[Brumby]

Curiosity got the better of me - and all I have is one of those resistance "wheels"  (The same as Jaycar carries).

$AUD30.41 (all up) for the 6 decade unit.  We'll see what I get (undoubtedly next year).

I had one of these bought some time ago, and to make it short: To anyone owning this decade box, put it in the trash, it’s the best advice I can give.

It may be ... but I'm OK with that risk.

[Brumby]

One thing that has always bothered me about these decade boxes is that they have a substantial offset; for example 0.4 which Dave zeroes out at the start.

But that means that if you want, say, 100.0 , you need to type in 99.6 . I can't help but think that there must be some clever way to incorporate this right into the mechanics of the box, such that if you ask for 0.0  , you get 0.4 , but if you ask for 100.0  , you get 100.0  . But I'm not clever enough to figure out such a trick. Is there a way? Other than using a microcontroller and 20 relays in a binary arrangement

That's problematic.

When you look at all the paths that can be taken, you would need to have that 0.4 reduction in every path - and ONLY ONCE in every path.  Certainly, you could ignore this in the higher ranges, but there is still the lower range or two, where the multiple path problem is stubbornly in your face.


In truth, this is just another example of the practicalities one must deal with in working in the real world ... with non-ideal components.

[PartialDischarge]

One thing that has always bothered me about these decade boxes is that they have a substantial offset; for example 0.4 which Dave zeroes out at the start.

That is solved by buying a good quality decade box.

[tggzzz]

One thing that has always bothered me about these decade boxes is that they have a substantial offset; for example 0.4 which Dave zeroes out at the start.

That is solved by buying a good quality decade box.

Yes, decent ones have a few milliOhms residual resistance. The lovely switches with springs and silver/graphite contacts can have a positively pornographic feel to them

There is now a fashion for calibrating resistance boxes to ignore the residual resistance, and only "count" the differential resistance. I dislike that intensely.

[Kleinstein]

If one does not like the switches, there are also cheap decade resistor PCBs that use jumpers to set the resistance. These seem to even a bit cheaper - though without a nice case.

[electr_peter]

Based on construction, description and price it seems that this decade resistance box is primarily intended for education purposes and not for regular lab use. Good value for money, but not comparable in quality to a proper unit.

How difficult would it be to redesign and replace whole PCB?

[AndyBeez]

I can see that very box turning up in a movie where the villains use it to crack the combination of a casino's vault door.

[PartialDischarge]

How difficult would it be to redesign and replace whole PCB?

The problem is not the PCB but the metal contacts, very bad quality, fail to make contact after some time.

[thm_w]

The problem is not the PCB but the metal contacts, very bad quality, fail to make contact after some time.

So was the metal losing springiness?
Or was there a conductive plating on them that gets wiped off.

[EEVblog]

Based on construction, description and price it seems that this decade resistance box is primarily intended for education purposes and not for regular lab use. Good value for money, but not comparable in quality to a proper unit.

Of course. No one expected that.

How difficult would it be to redesign and replace whole PCB?

Wouldn't be difficult, I actually thought about it.

[EEVblog]

One thing that has always bothered me about these decade boxes is that they have a substantial offset; for example 0.4 which Dave zeroes out at the start.

That is solved by buying a good quality decade box.

If you want the best:

[Neilm]

I'm guessing that it didn't come with a calibration certificate

[scopeman]

Well since we are talking decade boxes. I built this one in 1976. 1 ohm to 10megohm, 1 ohm steps 1% 1/4W resistors.
The switches are standard steatite Oak brand IIRC. It's built in a bakelite box. The cover was painted with blue Rustoleum paint and then a thin Mylar film was made and the labeling was done with Datack (SP?) rub on transfers. I had laid it on a drafting table with a paper layout behind it to help align the numbers. The overlay was put in place on the box and then a 1/16" sheet of Lexan was placed over that for protection.

Sometimes I wish it went to 11 on each decade 

The toggle switch switches between the resistance of the decade setting and the external resistor placed between the binding posts on the left. I was doing analog modem
design for a small company (first real job right out of school) and this feature made it handy for switching between mark and space frequencys all at 300 baud!

I still use it and it is still in spec.

Sam
W3OHM

[tggzzz]

Well since we are talking decade boxes....

Sometimes I wish it went to 11 on each decade 

It is good to homebrew

The good ones do go up to 11, of course - i.e. 0 to 10 on each decade 

The Cropico goes up to 11111ohms, the Muirhead to 1111.1ohms. The DekaStat even goes to 121.1ohms.

[foetusmachine]

Bought a couple of 4x units and modded them together to have 8 dials.

5w 1% resistors and the extra binding post is on order, but for the money, I think it's a pretty good project box.

[djQUAN]

I ordered the 5- and 6- knob versions when I saw the video and they just arrived. I modded both from ohms to kilo ohms so the 5 knob originally was 0.1 to 9999.9 is now 100R to 9.9999M
The 6 knob is now 100R to 99.9999M. more useful for my uses. I also slathered some dielectric grease on the contacts to hopefully extend the life for my very occasional use.

The slight offset is not much of a concern for substitution purposes in my case, if precision is required, I'll just measure the actual value then swap the nearest fixed resistor value after testing.

[foetusmachine]

My resistors arrived, but I forgot one value to make up the 500R value. So, now waiting for the last one before completing.

I do have a small offset, but that's measured with my hand held multimeter. I'll need to invest in something more precise for the low ohms to understand if/what adjustments may be requried.

[Kleinstein]

For 500 ohms one could use 2 x 1 K in parallel. 1 K shcould be there for the next decade.  I was anyway wondering why they did not use this to reduce the BOM.

[foetusmachine]

The original decade box used single, SMD resistors values that were "good enuf for china" , but well within the tolerance range - i.e. 49.9R for the 50R.

The board traces, contacts, terminals and wires are all going to add up to a more significant error here, particularly in the lower ranges.

In my case, I ordered packs of 5 5Wx1% resistors and I'd used four of these already in the 1k range, so I'm only left with a single 1k.

This is where i'm up to. I'll finish the soldering, clean up the board and then grease it all up for a bit more longevity.