Does Tolerance Matter In Crossover Capacitors

[thebes]

I want to freshen up my AK2 xovers with new caps, I'm considering Russian PIO's  which have a 10% tolerance, versus Dayton Audio Precision poly caps with a 1% tolerance.  Does tolerance make that much of a difference in xovers, cause they usually aren't that critical in things like tube amplifiers.

 

I thinking PIO because that's what PWK used when designing Khorns, and I liked them when I used them in my Cornwalls.

 

I'm also wondering why the 13uf used in the woofer is rated at 100volts, but the 3 2uf's for the tweeter are rated at 250volts.  Is that simply because that was what was available?

 

I'm also  wondering if anybody knows the voltage is for that 140uf cap for the bass bin, and if that's worth changing out.  All caps currently in there are circa 1984.

[jason str]

I have had good using the Dayton capacitors including the 1% type.

 

I'm no crossover expert but have updated my fair share of crossovers and there is no way you will be running 100 volts through them let alone 250 volts so it probably does not matter either way.

 

As far as tolerances go i was always told 10% is OK, its not like every driver is the exact same anyways.

[wvu80]

It is my understanding that the standard tolerances for the DA caps is 5%.  Then if you want to get the fancy 1% DA caps, they are the exact same ones except someone actually measures them and sorts out the ones that are within the 1% tolerance.  They cost more because there is an extra step involved.

[mark1101]

I always placed my orders by phone and asked them if they would measure the parts and match sets for me, and write the values on the caps.  I never had a problem when I did that and when I would curve the filter responses the left and right were on top of each other, dead nuts.  So I never played with the tolerances.

 

They must matter to some level of accuracy because you are creating specific filters.

[Quiet_Hollow]

They must matter to some level of accuracy because you are creating specific filters.

^^^ This.

 

I agree, and I would not totally discount the performance of a filter cap, considering the frequencies at play.

 

The purchasing department has the luxury of QA to weed out and match components. Us mortals OTOH only get what we get. So unless the plan is to buy 100+ components and sort through a pile, I think it'd be prudent to stick to the tightest tolerance available.

 

Unless of course, you can do as Mark suggests above. That'd be the way to go.

Edited February 4, 2016 by Quiet_Hollow

[thebes]

Well now I'm even more confused. That's never happened here before.

 

Why is Klipsch using such large caps with such high voltage measurements if they don't need to?  Why not just some 20volt caps?

 

In the past I ordered some from an ebay guy in Romania, sent them to Craig, who tested them  with a fancy cap tester and pronounced them worthy.

 

 

Doesn't surprised me that the 1% Dayton's are just culled from general runs, they used to do the same at Mullard with tubes.

 

Still I got nothing definitive yet. Do tolerances count in xovers, and what's the generally recognized voltage on replacement caps?

 

Inquiring Thebes-minds want to know.

 

 

 

 

 

I can understand the

[pzannucci]

As with any other components being replaced or when building a set of speakers, the tolerance doesn't matter (as long as you aren't way off) as much as matching.  I have heard tolerance is based on batching which won't help with singular caps.

 

To get good stereo imaging, the speakers have to match.  That means the more you like to see where the sound is coming from and speakers disappearing, the closer, at least in stereo, the right and left speakers need to match.  Bad part is drivers too which will be worse.

 

A small meter that can measure capacitance will be better than nothing here.  Bench measuring with load even better so you measure the whole crossover at one time to match the other.

 

So in ways tolerance does matter but not as much as component matching unless they are way off.

 

Also stay with the highest voltage you can get in your price range.  Nothing wrong with having headroom.  A lot of film types are much higher rated though cost much more but good for by-pass and are said to sound better.

[JJkizak]

Klipsch always wrote the actual measured values of the caps with crayon or something or other on each cap. At least they did on my 1965's.

JJK

[pzannucci]

Klipsch always wrote the actual measured values of the caps with crayon or something or other on each cap. At least they did on my 1965's.

JJK

Exactly my point 

[JohnA]

Yes, tolerance matters.  The closer to the design value you get, the closer the filter is to the frequency you wanted.  However, simpler networks are more tolerant, so, many of the AK-2 caps can be 10% tolerance, especially the 13 uF squawker cap.  The higher the "order" of the crossover, the more critical the tolerance.  Parts Express has a nice, affordable crossover design program that will let you play with the values and see the effect it has on the filter's response, if you want to get into the bug dust. 

 

Klipsch used the available motor start oil filled caps.  They were tough, cheap and did the job.  That is why the voltage is so high.  Think of it as putting V-rated tires on your Miata.  Stronger, but you'll never use it. 

 

I also check capacitance and match caps as much as practical. 

Edited February 6, 2016 by John Albright

[thebes]

Well I didn't play around  with the Parts Express xover program, but I did end up going with their 1% tolerance caps  for the mid-horn and tweeter, and poly caps for the four 70uf's dealing with the bass bin.  I may play around with Russian PIO's in the tweeter circuit. I have four of them here I bought for my Heresy's and never got around to installing.  I could do one set of PIO's for one speaker tweeter and compare the sound.

 

In the end, the cost difference between !% and 5% was negligible.

 

I suspect that down the road, I'll be putting all Russian PIO's in the mids and tweets.

[Deang]

If you're on a budget, hunt for the Russian PIO capacitors on eBay. Better yet, save for the Jensens or the Jupiters.

[mark1101]

I have 2 pairs of heresys that I had Bob Crites update the networks in years ago when he used to use the motor runs.  My son uses both pairs today and they sound great, honestly.  Back when Bob was using those, he was basically just taking out the old original motor runs that Klipsch used, and dropping in new ones.  The initial sound change was that you could hear the tweeter come back to life, and generally slightly clearer sound.  An improvement and a move back toward the original factory sound (probably).

 

However, at around the same time everyone got excited about the metalized poly's and film and foils and the so called boutique caps and how open they sounded, and how pretty they were in yellow.  .  That pretty much ended the game on the motor runs, as well as that they became scarce.

[Deang]

Bob was using the GE manufactured polypropylene in oil motor runs - Klipsch used Mylar (polyester) in oil and Paper in Oil. When GE ceased the manufacture of the motor runs, Bob couldn't find suitable replacements - so he went to the Sonicap - the only reasonably priced capacitor with a low ESR.

I preferred the Jensen paper in oils, but couldn't move the builds because of price, so I switched to Auricap and Kimber.

The best sounding capacitors in the old networks are the Jensen paper in oils and Jupiter wax/film. I still prefer the Type A and AA over anything else.

[thebes]

Dean is absolutely correct in his description of Bob's transition to Solen's. 

 

I was lucky enough to get some of his last motor run caps and they were the GE and are polypropylene.

 

Sorry to hear Dean that the cost of the new Jensens were so prohibitive. I betcha they sounded great. It's got to be either because of the costs of small production runs or outright greed on the manufactures part. Or both

 

I've got the Parts Express caps today and installed them.  Definitely smoothed out the tweets, maybe a little laid back in the bottom-end, but only time will tell.  An improvement for sure, but I never come to conclusions about specifics until some time has passed to give a sonic perspective.

[Mallette]

Why should you care if your capacitors want to crossover?  Times have changed and it's perfectly OK...

 

Dave

[DrWho]

Put on your engineering cap Thebes.....

 

Here's what the tolerance of a cap means:

If the cap is rated for 13uF with a 10% tolerance, then that means the capacitor you purchase will have a value somewhere between 11.7uF and 14.3uF.

A 13uF capacitor with a 1% tolerance will have a value somewhere between 12.87uF and 13.13uF.

 

So what does that mean for your crossover?

 

The crossover frequency is a function of the capacitor and inductor values. If you change the capacitor value, then the crossover frequency changes slightly. How much? Well it depends on the rest of the circuit.

 

At the end of the day, the tighter tolerance capacitor is guaranteed to be closer to the original design intent. That said, the value of the capacitors are random - so it's quite possible that your 13uF 10% capacitors come in closer to 13uF than your 1% tolerance caps. It really comes down to the luck of the draw.

 

 

In the case of a stereo pair of speakers, I would recommend sticking to 1% tolerance so that your Left and Right channels are more closely matched. Or you can buy a ton of 10% capacitors and then measure them with a cap meter and find a pair that match in value (one goes in the left speaker, one goes in the right).

 

 

 

As far as the voltage. When you exceed that voltage, the caps explode into a glorious fireball. Keep in mind though that a high Q filter can see more voltage across the caps than what the amplifier is putting out. Also, higher voltage caps tend to have less dielectric losses (the mechanisms for loss are typically the same mechanisms for voltage derating). Dielectric loss is bad because it "can" cause distortion....how much depends on the topology of the circuit, the type of cap, the voltage present, yadda yadda *yawn*.

[mustang guy]

I had to look it up, but 100 watts 8 Ohms is just over 28 volts.

 

It's the sq root of (watts times ohms).

[thebes]

Why should you care if your capacitors want to crossover?  Times have changed and it's perfectly OK...

 

Dave

Depends on where you crossover.  Case in point,. It all started in a titie bar in 1972.  She was named, Bill, she rode a horse named Jill, we danced and we played, we regretted and we paid. 

[Mallette]

Inadequate resistance, Thebes...  May she was really a transister and you were too charged up to notice.

 

Dave