Electrolytic vs film caps for smaller values.

[warpie]

..for low-value electrolytic capacitors you run into the problem with life expectancy correlating heavily with can diameter. meaning that physically narrow electrolytics are quite a lot more short-lived than their fatter cousins. For this reason, it's often advisable to use as high a working-voltage as there's physically room for in the circuit - and with recent miniaturization of PE-capacitors like the wima mks02 series it's simpler to go this route if you are designing for more-than-5years-service-life

/Jakob E.

example, e.g.:
https://www.chemi-con.co.jp/products/relatedfiles/capacitor/catalog/KZNLL-e.PDF

How low value is low? Do you mean anything below 100uF or anything below 470uF.

Also, if this is the case I would expect SMD 'lytics to last not long at all.

 

[57sputnik]

example, e.g.:
https://www.chemi-con.co.jp/products/relatedfiles/capacitor/catalog/KZNLL-e.PDF

If you look at the United Chemi-Con datasheet for the KZN-series you'll see (greyed out) that the 5, 6.3 and 8 mm can sizes are being discontinued.

I see a trend developing.

 

[abbey road d enfer]

..for low-value electrolytic capacitors you run into the problem with life expectancy correlating heavily with can diameter. meaning that physically narrow electrolytics are quite a lot more short-lived than their fatter cousins. For this reason, it's often advisable to use as high a working-voltage as there's physically room for in the circuit - and with recent miniaturization of PE-capacitors like the wima mks02 series it's simpler to go this route if you are designing for more-than-5years-service-life

/Jakob E.

example, e.g.:
https://www.chemi-con.co.jp/products/relatedfiles/capacitor/catalog/KZNLL-e.PDF

Then there will be no excuse for using low value coupling capacitors, based on the radio-age "10 times the value calculated for chosen LF cut-off".
The real estate cost of 10mm diameter vs. 5mm is partially compensated by the fact that more traces can be routed between pads.
Now it would be interesting to evaluate the various propositions for coupling in typical 10koms circuits.
270uF/35V vs. 120uF/63V vs. 56uF/100V, all of the same size, in circuits with bipolar 15-17V rails.

 

[Monte McGuire]

How low value is low? Do you mean anything below 100uF or anything below 470uF.

PMFJIH, but the can sizes like 5 mm, 6.3 mm and 8 mm are what I'd consider small. These are also in the realm where a film cap might be able to be substituted: e.g. a 1µF mylar is in the same basic PCB real estate size as a 1µF electrolytic in a 5 mm can.

100µF and 470µF caps usually have large enough cans that they do not seem to age prematurely.

Also, if this is the case I would expect SMD 'lytics to last not long at all.

That and the extended preheat and reflow times cause the part to reach "really high" temperatures, which seems like a cruel thing to do to an electrolytic. I'm designing a device that uses a select few large electrolytics, but I'm going the other direction - soldering in Mill Max sockets so I don't have to solder the through hole caps at all - no solder heat torture to start their life. Will make some future tech's recap a breeze - trim the leads and plug the new caps in. These receptacles seem to perform well too.

 

[Tubetec]

'I'm taking no risks, I always aim to the less distortion possible'

Modern op-amp based circuits tend to work very well upto their max output , then distortion rises very quickly and the whole thing clips in a very non musical fashion . In real world situations we can never anticipate precisely the SPL thats going to come from the mouth of a vocalist during a performance. Equipment that clips more gracefully has long found favour with sound engineers , were taking tubes, transformers , discrete class A transistor circuits , ok it may not match the low distortion possible with modern designs , but on the peaks you have some extra insurance against things getting really nasty . Back in the days before limiters and compressors designers built stuff that distorted gently to help prevent overloading the transmitters , the early BBC designs exemplify this approach .

Interesting idea Monty , had to ask what does PMFJIH mean ?

 

[Monte McGuire]

Then there will be no excuse for using low value coupling capacitors, based on the radio-age "10 times the value calculated for chosen LF cut-off".
The real estate cost of 10mm diameter vs. 5mm is partially compensated by the fact that more traces can be routed between pads.

I like this thinking...

Now it would be interesting to evaluate the various propositions for coupling in typical 10koms circuits.
270uF/35V vs. 120uF/63V vs. 56uF/100V, all of the same size, in circuits with bipolar 15-17V rails.

If one trusts the data sheets, one difference among these choices will be that the tan-theta specs for most caps depends solely upon the rated voltage, and not on a CV product. So, to reduce the tan-theta losses, which are basically the real component of the dielectric losses, you want the higher voltage rating. Again though - it would be good to see if that ends up being an issue or not re. distortion or "sound quality".

I'd also imagine that leakage could be different among these choices, as leakage is usually specified as a multiple of the cap's CV product. Still, that's "datasheet spec land" and it would be interesting to see how the parts actually behave.

 

[warpie]

Interesting idea Monty , had to ask what does PMFJIH mean ?

Pardon Me For Jumping In Here

I had to google it

 

[warpie]

That and the extended preheat and reflow times cause the part to reach "really high" temperatures, which seems like a cruel thing to do to an electrolytic. I'm designing a device that uses a select few large electrolytics, but I'm going the other direction - soldering in Mill Max sockets so I don't have to solder the through hole caps at all - no solder heat torture to start their life. Will make some future tech's recap a breeze - trim the leads and plug the new caps in. These receptacles seem to perform well too.

If I understand correctly you mean that you'll solder the Mill Max on the PCB and then simply "snap in" the capacitors? I think this approach is looking for troubles. I can foresee loose contacts due to shipping, oxidation, etc.

Maybe I totally misunderstood what you wrote...

 

[Monte McGuire]

If I understand correctly you mean that you'll solder the Mill Max on the PCB and then simply "snap in" the capacitors? I think this approach is looking for troubles. I can foresee loose contacts due to shipping, oxidation, etc.

No, you got it right, and yes, one needs to find an alternate system for mechanically anchoring the capacitor to the PCB, so that it does not vibrate and cause 'fretting corrosion' to the contacts.

It's maybe overkill, but in this device, the only elcaps are power supply isolators - exactly one cap value applied in multiple places, and in my prototype PCBs, I decided to use the Mill Max receptacles sized for the elcap leads, and they performed so well and didn't cost that much, so I decided to just make that the norm. The cool thing is that I can reflow the Mill Max parts easily with the rest of the circuit, and not have to do a hand solder (post reflow) of an elcap into a 4 layer board, which is always a little creepy due to the large amount of copper in my PCB layouts - these really want to be reflowed and not hand soldered. Plated through holes are always suspect, but if they are reflowed with solder paste and a Mill Max socket, then I tend to trust them more than hand soldered elcaps.

 

[Tubetec]

Cheers Warpie !

Maybe a nylon P clip of the appropriate diameter would be a simple and cheap way to fasten the caps down to deck .
The extra lead lenght might reduce stresses within the cap itself and the solder joints .

 

[abbey road d enfer]

I'm designing a device that uses a select few large electrolytics, but I'm going the other direction - soldering in Mill Max sockets so I don't have to solder the through hole caps at all - no solder heat torture to start their life.

I did something similar some time ago, not for this reason. The thing is the client wanted to select capacitors subjectively. The big issue is that each capacitor seems to have its individual lead gauge...

 

[JohnRoberts]

If you look at the United Chemi-Con datasheet for the KZN-series you'll see (greyed out) that the 5, 6.3 and 8 mm can sizes are being discontinued.

I see a trend developing.

I don't know if these two trends are related but I recall the difficultly sourcing cheap, linear, SMD caps a decade or two ago. I had one production run of promising looking SMD film caps that melted during reflow.

NPO/COG SMD caps delivered excellent audio performance but were generally not available in larger values. A quick glance over at DIGIKEY just now reveals 0.01uF SMD NPO/COG caps so this can handle pretty much all audio filter needs.

DC blocking applications may remain electrolytic, but if you can hear an electrolytic in this application you are doing it wrong.

JR

 

[abbey road d enfer]

I'm designing a device that uses a select few large electrolytics, but I'm going the other direction - soldering in Mill Max sockets so I don't have to solder the through hole caps at all - no solder heat torture to start their life.

I did something similar some time ago, not for this reason. The thing is the client wanted to select capacitors subjectively. The big issue is that each capacitor seems to have its individual lead gauge...

 

[JohnRoberts]

That and the extended preheat and reflow times cause the part to reach "really high" temperatures, which seems like a cruel thing to do to an electrolytic. I'm designing a device that uses a select few large electrolytics, but I'm going the other direction - soldering in Mill Max sockets so I don't have to solder the through hole caps at all - no solder heat torture to start their life. Will make some future tech's recap a breeze - trim the leads and plug the new caps in. These receptacles seem to perform well too.

Modern capacitors are designed to survive reflow ovens (just like wave soldering before). Except for that batch of SMD film caps I encountered some 20 years ago that literally deformed while melting at my contract manufacturer.

Heat damage to electrolytic capacitors is cumulative over time and in my judgement dominantly loss of electrolyte. Cap makers sell capacitors rated for higher ambient temperature use. They probably use different bungs and may begin with more electrolyte but likely are otherwise similar.

JR

 

[Monte McGuire]

I did something similar some time ago, not for this reason. The thing is the client wanted to select capacitors subjectively. The big issue is that each capacitor seems to have its individual lead gauge...

Fortunately, these caps are largely similar to each other. For 10 mm diameter cans, 0.6 mm leads are almost universally used.

 

[abbey road d enfer]

Fortunately, these caps are largely similar to each other. For 10 mm diameter cans, 0.6 mm leads are almost universally used.

Right, but it'll take some time for replacing my inventory with "standardized" 10 mm types.
Until recently, I did all my PCB layouts with 5mm spacing, even for the smaller types, which implied some lead bending.
Then, in a burst of consciousness, I decided to do the right thing, and used, 2.5 and 3.5mm spacing whenever necessary.
Now, I have to revert to my old habits...

 

[user 53957]

I agree regarding distortion, not so much about phase response.

What do you mean? Do you have a better method for transmitting signals from one stage to another? Do you suggest using servos around all stages? Or transformers?

Phase response fallows the transfer function of the coupling circuit. In certain applications, a designer will look at this. In small signal amplification designs, not much attention is paid to this other than measuring the effect afterwards for bench marking on a datasheet.

The capacitor is the cheapest method of AC coupling. Cheapest as in the cost of the part itself, and design considerations involved in compared to the better, but more expensive method called transformer coupling. Transformer coupling is superior becauese the designer can make the transformer specs in circuit exceed the bandwith response curves the capacitor circuits can achieve.

But the best method of coupling is always DC coupling because it does not limit bandwidth nor add distortion, besides self noise in resistors, that can be selected for better noise profile.

 

[abbey road d enfer]

Phase response fallows the transfer function of the coupling circuit. In certain applications, a designer will look at this. In small signal amplification designs, not much attention is paid to this other than measuring the effect afterwards for bench marking on a datasheet.

You have a very strange view on how products are designed. Of course, any half-decent designer will determine the value of coupling capacitors, according to the surrounding circuit's impedances and a targetted response.

The capacitor is the cheapest method of AC coupling. Cheapest as in the cost of the part itself, and design considerations involved in compared to the better, but more expensive method called transformer coupling.

It is known since ages that transformers are very unperfect, with limited frequency response and distortion issues. For many years, it was just about the only option in some circuits, but today, there is no objective reason to use a transformer in an audio circuit. Today ransformers are used for a supposed euphonic quality, for respect of a vintage circuit or for marketing reasons.
With solid-state electronics, there is no circuit where a transformer could not be replaced with a good capacitor.

Transformer coupling is superior becauese the designer can make the transformer specs in circuit exceed the bandwith response curves the capacitor circuits can achieve.

It is extremely difficult to increase the frequency response of a transformer, because for increasing the LF response, the inductance must be increased, which results in increased leakage capacitance, which in turn reduces the HF response. In other words, a compromise must be found between LF and HF extension.
Conversely, increasing the value of a coupling capacitor, for lowering the LF cut-off, results in almost no effect on the HF response.

But the best method of coupling is always DC coupling because it does not limit bandwidth

Is it really necessary, or even desirable? What's the point of passing 1Hz at 0dB, when no transducer can make any use of it?

nor add distortion, besides self noise in resistors, that can be selected for better noise profile.

How can DC coupling improve noise?

 

[user 53957]

You have a very strange view on how products are designed. Of course, any half-decent designer will determine the value of coupling capacitors, according to the surrounding circuit's impedances and a targetted response.

I come from a broad electronics background. So designing audio circuits, much less analog circuits would be very nostalgic to me. But If you want to trade notes on mathmatical formulas used in coupling circuits and have a true scientific discussion about this you are wlcome to start another thread on your site particularly for this.

It is known since ages that transformers are very unperfect, with limited frequency response and distortion issues. For many years, it was just about the only option in some circuits, but today, there is no objective reason to use a transformer in an audio circuit. Today ransformers are used for a supposed euphonic quality, for respect of a vintage circuit or for marketing reasons.
With solid-state electronics, there is no circuit where a transformer could not be replaced with a good capacitor.

You need to unlearn that. Transformers can be engineered however an EE wants them designed. I remember someone asking me once my advice on transformer coupling designs at one of my lectures that I gave over RF transformer core constructions, later, I found out that man was one of the engineers at Jensen and about a decade or so later, he made a series of high end interconnect transformers using the standard industry 8-12MHz core construction techniques I lectured on.

A quick glance at their site revealed one that DIN-PB that he designed.

I'm sure he has others as well, and of course, there are other guys out there that can build just as good.

Maybe your view of transformers was limited to the splattering of the mid-low end UTC transformers audio transformers of the past?

Transformers are still used today and in even different designs. From power supplies to network switches, and routers, to electricity power meters. It is not my fault low end audio only use them as 'audio gimmicks' with inexpensive transformers.

How can DC coupling improve noise?

I never said anything about improving noise. Resistors can be selected for the best low noise profile while offering low insertion loss when DC coupling. The trick in all of this is to do this outside the operating range so the design does not alter the signal in an undesirable way.

 

[57sputnik]

How can DC coupling improve noise?

In a mic preamp or moving coil phono preamp the input coupling capacitors' reactance allows low frequency input noise voltage to develop across the capacitor(s). The reactance(s) appear in series with the source impedance.

Larger values of course minimize this but phantom fault stored charge limits practical input values to 100 µF or less for mic preamps.