Built the Wallin 2 jig using parts from Dick Smith Electronics (a RadioShack like supplier in Aus/NZ).
Problem is the 16 / 4 ohm callibration resistor impedance when measurements are made between BP3 and BP4.... is approx 31 ohms and 14 ohms respectively. (Doing the last 2 tests listed on Eric's site to verify jig operation).
All my other measurements are out too - but when you are testing the line level reducing circuits (ie. the 1.175K ohm ones) - then being out 20 ohms doesn't matter.... but for the small resistances, it makes a large margin of error!
I can put some of it down to series resistance and out of tolerance callibration resistors (even though their are metal film 1% precision types).
I found measuring impedance between switch posts ranges from 1.5ohm to 2.5 ohms depending on the switch!. When I get home tonight - I'll post all the individual series resistances for each leg of the circuit to try and pinpoint the problem.
Have I got some faulty switches? Is there anything that can be done to reduce resistance on them? (ie. is it internal lubrication, bad contacts etc...). Do I need to fix this? (presume if I don't it will lead to wildly inaccurate speaker impedance measurements)
Any help appreciated.
Thanks,
David.
Problem is the 16 / 4 ohm callibration resistor impedance when measurements are made between BP3 and BP4.... is approx 31 ohms and 14 ohms respectively. (Doing the last 2 tests listed on Eric's site to verify jig operation).
All my other measurements are out too - but when you are testing the line level reducing circuits (ie. the 1.175K ohm ones) - then being out 20 ohms doesn't matter.... but for the small resistances, it makes a large margin of error!
I can put some of it down to series resistance and out of tolerance callibration resistors (even though their are metal film 1% precision types).
I found measuring impedance between switch posts ranges from 1.5ohm to 2.5 ohms depending on the switch!. When I get home tonight - I'll post all the individual series resistances for each leg of the circuit to try and pinpoint the problem.
Have I got some faulty switches? Is there anything that can be done to reduce resistance on them? (ie. is it internal lubrication, bad contacts etc...). Do I need to fix this? (presume if I don't it will lead to wildly inaccurate speaker impedance measurements)
Any help appreciated.
Thanks,
David.
It's hard to make accurate measurements in the 4-8 ohms region with many switches. You either need switches with high spring pressure, designed for a lot of current, or gold contacts like the little Alcoswitch or other units. Decent ceramic wafer switches can also do it, but that doesn't help you here. If you've ever seen an old style resistor box or low resistance bridge, you'll see large contacts that operate with multiple wipers and lots of spring pressure, or even brass taper plugs. Sorry, but I think you need better switches- I don't know of a fix for the ones you've got.
One option I have is replacing the switches with RCA connector pairs to have a semi-hardwired solution. At least that would minimise contact resistance although it would:
a) cost a lot more (RCA "jumper" cables to plug between the virtual "switch" settings plus RCA female sockets - 2 pair for SW1 and another 2 pair for SW3)
b) be a hassle having to plug / unplug
David.
a) cost a lot more (RCA "jumper" cables to plug between the virtual "switch" settings plus RCA female sockets - 2 pair for SW1 and another 2 pair for SW3)
b) be a hassle having to plug / unplug
David.
Just taking a quick look on-line, Digikey has C&K DPDT toggles under p/n CKN1089-ND. Don't be shocked- they're about $9 each. They spec about 20 milliohms initial contact resistance for both silver and gold contacts. Possibly you can find something similar surplus or overstock for less. I wouldn't use RCA connectors, as they're highly variable and not really designed for low resistance contact. Decent quality banana plugs and jacks are pretty good, but it seems like it would take a lot of them, and be a pain to use.
I built a Kelvin-Varley divider once, and had the same issue- what to use for a reliable low resistance switch. I ended up buying a Molex terminal header (row of pins) having a big spacing (.156"), and some matching Molex connector bodies and pins. I wired the connector bodies as jumper blocks, then just plugged them onto the header in the positions I needed. It worked well, but not as elegant as decent switches. OTOH, I was able to work at the 5 parts per million level for just a few dollars, so I was reasonably pleased.
I built a Kelvin-Varley divider once, and had the same issue- what to use for a reliable low resistance switch. I ended up buying a Molex terminal header (row of pins) having a big spacing (.156"), and some matching Molex connector bodies and pins. I wired the connector bodies as jumper blocks, then just plugged them onto the header in the positions I needed. It worked well, but not as elegant as decent switches. OTOH, I was able to work at the 5 parts per million level for just a few dollars, so I was reasonably pleased.
Thanks Conrad! I have one other question: I'm building an A/B switcher for speaker auditioning and I need high quality potentiameter (sp?) to set the levels equal. Is there anything high quality at Digikey for this? I really don't know what makes a good "switch" so I've held off on making this box. thanks again for the help.
I don't really know what makes a really good potentiometer- I like the big old Allen Bradley carbon pots, but that's more for ruggedness than any signal passing property. To do a valid A-B test, you need to match levels with good precision, 0.1 dB worst case I believe. You might consider a two-knob control. Use a dual wafer rotary switch and say 12 steps using good metal film resistors, then add a pot for fine control. That would take some of the burden off the pot for good matching between the sections. There's probably some good arrangement to do this, but I'll have to think about it. A neat way would be similar to the above mentioned Kelvin-Varley technique, but it would take a four-wafer switch to do two channels properly.
Thanks for the info Conrad.
Unfortunately, shipping from Digikey is too expensive to NZ.
Last night I "worked" the switches many times and got the 4 / 16 ohm settings down to 6 and 18 ohms respectively. This however is not consistent (I still had +/- 15 ohm variations on both circuits) and would have to be re-measured and factored each time I do an impedance measurement (a real headache).
Since I can't source low resistance switches here, I might just have to go the hard wired route. The other option might be to lobotomize a defunct vintage receiver with those big old heavy switches they used to use.
Dave.
Unfortunately, shipping from Digikey is too expensive to NZ.
Last night I "worked" the switches many times and got the 4 / 16 ohm settings down to 6 and 18 ohms respectively. This however is not consistent (I still had +/- 15 ohm variations on both circuits) and would have to be re-measured and factored each time I do an impedance measurement (a real headache).
Since I can't source low resistance switches here, I might just have to go the hard wired route. The other option might be to lobotomize a defunct vintage receiver with those big old heavy switches they used to use.
Dave.
It's hard to believe those switches are that bad. Maybe they have too much lubricant. Maybe drill a small hole in the side and flush them with solvent? Do you have a hardware store that sells big toggle switches for the mains? If you switched a motor or large bulb on, and had a couple ohms switch resistance, the switch would burn up! If you can't get switches, the pin header route would work well. There must be some reasonably available Molex distributor around.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.