This coming from one who has matched hundreds of LDR's, maybe a thousand or more, it sounds interesting Ri, would you like to share it?
Probably around 4-5k.
Sure.
Its pretty simple. So you have two LDRs. They are either series or shunt. But both the same. Send positive voltage to both + leads of the LDRs. You need two regular resistors and a trimmer. You COULD use only a trimmer but it is to dramatic so the regular resistors help. You will see. You take one regular resistor to the negative lead of each LDR. Then connect each of the regular resistors to one lead on the trimmer. Connect ground or negative voltage to the wiper. When you turn on the LDRs and measure them you can dial the trimmer for more or less resistance on each. But what happens is the resistance 'pans' back and forth so while one decreases in resistance the other increases. It seems to be a better solution than putting one single trimmer in series with the positive lead of the loud channel which has been the standard so far. The LDRs, with this 3 resistor way, seem to stay in a more parallel relationship. If you were to graph them their lines would be smack on top of each other for full volume pot rotation.
This little fix should be done in most cases on series LDRs. And I would do it with the system running rather than with a DMM. WHY? Because if its your shunt LDRs that are screwed up and you are trying to correct the problem by addressing your series LDRs then your DMM wont tell you anything that will immediately make sense, but your ears will. See, LDRs that match at 1k, 4k, 10k for example will probably not match at 200R and your shunt LDRs will need to match really well at low resistance but thats the hardest area to get LDRs to match. So lets say you worked on getting the shunt to match with this little circuit at the area where they dont match. Okay you bring one from 150R to 200R and the other from 250R to 200R. Great! They match! But at this resistance they are sucking more power than at any other setting and they are greedy so they are not very predictable here. Way more predictable at higher resistance. Even up at 500R they make way more sense and area already acting at 500R they way they will be acting at 1k, 4k, 10k. So if we force 150R and 250R to be 200R then we have changed the way they react to power only when they are suck hogs and when they get into higher resistances we are not letting them act they way they naturally would so it will result in a worse match. But if we force the series resistors to do the compensation for the shunts we are doing this at the highest resistances the LDRs will ever be at and the series will never be down at suck hog resistances so we are only nudging them and they will accept that nudge and continue to have a resistance curve that is not bulged out at one point but just nudged over to bring the channel into balance and it will keep its natural curve. What we match is the natural curve anyway so keeping the curve means that even though we nudged the channel over a little it will still increase/decrease in volume in a way similar to the other channel. I really talk to much.
So since I have said do this on Series LDRs to compensate for series OR shunt mismatch I am going to say try it on shunt to. WHY? Cuz shunt are suckhogs as they are most often in a low resistance situation where LDRs are suckhogs. However, you might get lucky and they suckhog the same way. For instance if you give both 100mW do they both hit 50R? If they do then concentrating on shunt to get a more balanced volume is great and probably more effective. This is assuming they are a match up in that 1k, 4k, 10k. If they dont match at all then its not a match and it wont be a match and no analog circuit is going to make them match.
I have a circuit board to control LDRs that lets me change min and max resistance as I please which has been really nice to have in the DCB1. See this isnt OT at all
Its my prototype right now but I will have it available if anyone wants them in the next few weeks. This little balance circuit is something I will include in the final copy of the board. It will actually be on a separate board where the LDRs will live and there will be two wires from the power board to the LDR board. This way you can mount the LDRs right on the RCA input/outputs and not have the power supply right on top of the injecting noise into the signal.
Different impedances sound different. I prefer lower impedances as they sound more transparent and bright. So I like to turn mine down to 6k as I have said. The problem with LDRs in the DCB1 circuit or any other is that no matter what taper you use on your control pot the LDRs have a taper of their own and when the taper of the control pot gets steep and the LDR's taper starts to get steep at the same time, watch out, the volume increase is almost to fast to handle. So with this board in my DCB1 I can tell the series LDRs that they will operate over, for instance, a range of 4k-8k.... I can tell the shunt to operate from 50R to 2k. So when I use a single control pot with dual gangs on it then the taper of the pot is much more a good thing then a bad thing because the only LDRs that even come close to their sharp increase in resistance is the series while the shunt are quite linear at this point. So my volume increases slowly and I get full rotation of my pot. I can turn the pot up to max rotation and then dial up or down my LDRs on the control board giving me the ability to choose my max volume. Now the volume will only be as loud as I want it to be at max pot rotation. I can do the same for min volume.
Now this is obviously off topic and it makes me think that a thread that answers questions about how LDRs work and how to control them would be a good idea.
Salas, while this information is kind of relevant to the DCB1 because some people are building with LDRs I do realize its kind of OT and I wont go on about it anymore. Thats why I tried to throw everything I could think of into one post.
Ri
Sure.
Its pretty simple. So you have two LDRs. They are either series or shunt. But both the same. Send positive voltage to both + leads of the LDRs. You need two regular resistors and a trimmer. You COULD use only a trimmer but it is to dramatic so the regular resistors help. You will see. You take one regular resistor to the negative lead of each LDR. Then connect each of the regular resistors to one lead on the trimmer. Connect ground or negative voltage to the wiper. When you turn on the LDRs and measure them you can dial the trimmer for more or less resistance on each. But what happens is the resistance 'pans' back and forth so while one decreases in resistance the other increases. It seems to be a better solution than putting one single trimmer in series with the positive lead of the loud channel which has been the standard so far. The LDRs, with this 3 resistor way, seem to stay in a more parallel relationship. If you were to graph them their lines would be smack on top of each other for full volume pot rotation.
This little fix should be done in most cases on series LDRs. And I would do it with the system running rather than with a DMM. WHY? Because if its your shunt LDRs that are screwed up and you are trying to correct the problem by addressing your series LDRs then your DMM wont tell you anything that will immediately make sense, but your ears will. See, LDRs that match at 1k, 4k, 10k for example will probably not match at 200R and your shunt LDRs will need to match really well at low resistance but thats the hardest area to get LDRs to match. So lets say you worked on getting the shunt to match with this little circuit at the area where they dont match. Okay you bring one from 150R to 200R and the other from 250R to 200R. Great! They match! But at this resistance they are sucking more power than at any other setting and they are greedy so they are not very predictable here. Way more predictable at higher resistance. Even up at 500R they make way more sense and area already acting at 500R they way they will be acting at 1k, 4k, 10k. So if we force 150R and 250R to be 200R then we have changed the way they react to power only when they are suck hogs and when they get into higher resistances we are not letting them act they way they naturally would so it will result in a worse match. But if we force the series resistors to do the compensation for the shunts we are doing this at the highest resistances the LDRs will ever be at and the series will never be down at suck hog resistances so we are only nudging them and they will accept that nudge and continue to have a resistance curve that is not bulged out at one point but just nudged over to bring the channel into balance and it will keep its natural curve. What we match is the natural curve anyway so keeping the curve means that even though we nudged the channel over a little it will still increase/decrease in volume in a way similar to the other channel. I really talk to much.
So since I have said do this on Series LDRs to compensate for series OR shunt mismatch I am going to say try it on shunt to. WHY? Cuz shunt are suckhogs as they are most often in a low resistance situation where LDRs are suckhogs. However, you might get lucky and they suckhog the same way. For instance if you give both 100mW do they both hit 50R? If they do then concentrating on shunt to get a more balanced volume is great and probably more effective. This is assuming they are a match up in that 1k, 4k, 10k. If they dont match at all then its not a match and it wont be a match and no analog circuit is going to make them match.
I have a circuit board to control LDRs that lets me change min and max resistance as I please which has been really nice to have in the DCB1. See this isnt OT at all
Its my prototype right now but I will have it available if anyone wants them in the next few weeks. This little balance circuit is something I will include in the final copy of the board. It will actually be on a separate board where the LDRs will live and there will be two wires from the power board to the LDR board. This way you can mount the LDRs right on the RCA input/outputs and not have the power supply right on top of the injecting noise into the signal. Different impedances sound different. I prefer lower impedances as they sound more transparent and bright. So I like to turn mine down to 6k as I have said. The problem with LDRs in the DCB1 circuit or any other is that no matter what taper you use on your control pot the LDRs have a taper of their own and when the taper of the control pot gets steep and the LDR's taper starts to get steep at the same time, watch out, the volume increase is almost to fast to handle. So with this board in my DCB1 I can tell the series LDRs that they will operate over, for instance, a range of 4k-8k.... I can tell the shunt to operate from 50R to 2k. So when I use a single control pot with dual gangs on it then the taper of the pot is much more a good thing then a bad thing because the only LDRs that even come close to their sharp increase in resistance is the series while the shunt are quite linear at this point. So my volume increases slowly and I get full rotation of my pot. I can turn the pot up to max rotation and then dial up or down my LDRs on the control board giving me the ability to choose my max volume. Now the volume will only be as loud as I want it to be at max pot rotation. I can do the same for min volume.
Now this is obviously off topic and it makes me think that a thread that answers questions about how LDRs work and how to control them would be a good idea.
Salas, while this information is kind of relevant to the DCB1 because some people are building with LDRs I do realize its kind of OT and I wont go on about it anymore. Thats why I tried to throw everything I could think of into one post.
Ri
In this case I'd rather fix both the mosfets and the pcb to the roof. I'd like to use a little heatsink for the mosfets that will stay on top of the case. Removing the heat through the side walls is not fesible for me because the case will sit very close to a piece of furniture so only the top will be ventilated enough.
What means "very close", is the length of the resistors' legs (say, 3-4 cm) close enough or you really mean very close, as in <1cm between the body of the mosfet and the body of the resistor? Sorry, I hope I'm not cluttering this thread with my newbie questions, but it seems the best place to ask them.
Tea-Bag said:
You would do what, floor or roof?
So I ordered a bunch of LEDs from Mouser. I got the correct part numbers. But it seems all of the LEDs measure around 1.945 or so. That makes it imposable to get 5.4v for three and 9 volts for 5.
So... may I get some 1.7 and mix and match that way?
Dumb *** question but I shall wait with baited breath for a reply.
So... may I get some 1.7 and mix and match that way?
Dumb *** question but I shall wait with baited breath for a reply.
One quick question, please. Is there a problem if the IRFPs will be a few centimeters (no more than 8-10) away from the board? I am considering mounting them on the roof (of the case
You can mount them quite a distance if you need to.
Of course you don't really want long wires which might
pick up noise...
I got the same problem. What's the part number for the leds with 1.7 forward voltage drop? (and where did you get them?)
If someone reads behind I have expanded enough times. Everything around 9.5-11V for output voltages will do OK. 10 is the slightly subjectively preferred center. The positive side should end up slightly more, due to more current through its Leds. Just match the strings per side. You can always mix the odd weaker Led if your mean average is stronger than you might have liked. For the Njfets 0.1mV Idss match will do.
DACT Type 21 Stepped Attenuator 20K FOR PASS B1 DIY - eBay (item 120502791204 end time Jan-05-10 09:17:39 PST)
How about these? I just bought two.
How about these? I just bought two.
Last edited: