Hi Ida, I sent you a link which explains all of it. Room EQ Wizard is a free speaker / room measurement software which has been enhanced to also do impedance measurements. This includes speakers, caps and coils. It has a vibrant following, and you'll need to ask their discussion groups for more answers, but here is the link to the software: https://www.roomeqwizard.com/And in particular, here is the description of how to do impedance measurements with it: https://www.roomeqwizard.com/help/help_en-GB/html/impedancemeasurement.htmlBut I don't actually use it, I use low-end hobbyist/pro gear instead. Still, this would be among your cheapest options. Best, E |
I'm not able to read carefully, so I may not be understanding correctly, but I think the issue is that Tim hasn't seen this problem before. :) I have, and I've gone through the modelling. This would be the third or fourth time I've seen this issue in total, and the fix is always "magic." :)
What no one believes until they see it is that the problem caused by the low ESR tends to not affect the section the caps are in, but the next one up. That makes it maddening to trouble shoot. People spend all their time and effort trying to fix the HP tweeter section, but the problem is caused by the mid or woofer section (whichever is next down). In a 3-way, these impedance issues can be compounded.
The issue isn't really whether you use polar or non-polar. It's the total ESR that is playing games with you. Tim is quite right that having 2 caps in series is going to double the ESR, so maybe that one could need an extra 2-4 Ohms.
With a little hacking of a couple of audio cables you can make your own cap measurement jig and use Room EQ Wizard. That will be pretty accurate and cheap.
Best,
E
|
Hi Ida,
I'm not advocating you keep all the old caps. Just the one's in parallel with the drivers. Those "shunt" caps.
So long as you are having a good time and learning new things, have at it.
What you might also enjoy, if you get Room EQ Wizard, is measuring your system impedance with the caps. Also measure the individual driver impedance (in box)
Then you can use something like XSim to simulate most of it and play around with different cap ESR in simulation. This will really help you see what is going on.
Best,
E
|
Actually, @idahifi, that's my other bit of advice.
Restore the original shunt caps and leave it alone. :)
|
Idahifi's experience is exactly what I've been talking about. :)
In a 2nd order hp filter, this problem occurs from a combination of:
- Very low DCR in the coil
- Very low ESR in the caps
- Too small a resistor in series with the cap
- The impedance of the next stage higher in Hz.
In the right combination of things, you end up with far too low an impedance in the next stage.
Of course, having a monstrous amp fixes this issue, but you'd be surprised how just a little dip around 3 Ohms or less can do this.
Inexperienced modders get themselves into trouble by trying to get the lowest possible coil DCR and cap ESR without understanding the entire circuit, and bam.
Best for them to only replace series caps. There may be some level shifting but it is more benign.
Best,
E
|
Idahifi,
Unless otherwise specified, electrolytics are polar, but crossover caps must all be bi-polar (same as non-polar). :)
You can "cheat" by putting polar caps back-to-back, turning 2 polar into effectively a single non-polar with half the effective capacitance.
Film caps as far as i know are all bi-polar.
The rare exception to this is with battery biased crossovers. Very very rare. :)
Best,
E
|
|
Idahifi,
no. ESR = equivalent series resistance.
It is measured at specific frequencies, and may be different.
Again, i don't know why you are going down this route. Experiment first, replace the caps I listed. See if this fixes things. THEN if that is the problem you'll need to get measurements.
Best,
E
|
idahifi,
Consistent with my posts.
Best,
E
|
Hi Tim,
Yes, your analysis is correct. :)
Too low impedance in the second poles _may_ lead to lowering of system impedance above the cut off. I know it’s weird, but I’ve seen it and modelled it a few times. It is a real effect.
Whether that is what the OP is suffering from I am not sure, but the series resistors are a clue. Usually those are in place to prevent a dead short. Fortunately, the proof of this hypothesis is cheap. :) Just replace the caps with their originals and listen. If that proves to be correct, then the OP knows what to fix.
Best,
E
|
So looking at the schematic, the two caps I would be most suspect of are C3 and C4. Note they have a resistor in series with them, so the overall resistance there must be important. I would suggest you go back to the original caps there, and listen again.
If this fixes your problem, you’ll want to figure out the ESR of each one by substituting larger R3/R4 to make up the differences.
Almost missed it, C1 is the same way. My suspicion is that the resistors in series are going to be really small , like 2-5 Ohms max.
Best,
E
|
Idahifi, Bring your problem over to DIYhifi, where you will find tons of speaker design geeks in the Multi-Way forum. :) http://www.diyaudio.com/forums/multi-way/I can help you there more easily with pics, diagrams etc. Best, E |
Hi Idahifi,
I think I know what happened. You removed electrolytic shunt caps without taking into account the reduced ESR.
The easiest way to fix this is to replace the original shunt caps. By shunt, I mean caps that go towards ground, and are not in series with the driver.
If you have the right measurement tools, like DATS or REW with the right jig, you can measure the original caps, and compensate for the low ESR with added resistance in series.
It is particularly difficult to find this issue for most because the problems tend to occur where you don't expect it. That is, the treble droop happens because of something you did in the midrange filter circuit.
This problem also occurs with series caps, but changes there tend to be much more subtle, it's hard to go entirely wrong. The shunt caps, having the ability to create a short to ground, are the most troublesome.
Best,
E
|
