If construction quality is not taken into account, perhaps polystyrene is the best, but to really get that good the cap MUST be constructed well. Usually they aren't. That is my understanding at least.
Conrad Hoffman has made some interesting tests using a bridge on a variety of caps, available at his site. I'm not sure I would trust styrene caps too much. One of the ones I have has loose leads which can be pulled right out and inserted back again. Others have tears in the outer film, exposing the foil. If you are serious, I think you would have to test them yourself.
There are silver mica SMD caps on Mouser, even Teflon caps from 1pF-9pF.
- keantoken
Conrad Hoffman has made some interesting tests using a bridge on a variety of caps, available at his site. I'm not sure I would trust styrene caps too much. One of the ones I have has loose leads which can be pulled right out and inserted back again. Others have tears in the outer film, exposing the foil. If you are serious, I think you would have to test them yourself.
There are silver mica SMD caps on Mouser, even Teflon caps from 1pF-9pF.
- keantoken
Yes, Mike, I do. PS is good, but silver mica is better still - but expensive.
HOSFELT ELECTRONICS, INC., in Steubenville, OH.
Cheers,
Hugh
HOSFELT ELECTRONICS, INC., in Steubenville, OH.
Cheers,
Hugh
I have never seen such low DF on a datasheet. Usually low-loss caps will specify DF in the datasheet. I've only seen as low as .03% for the Evox Rifa PFR series. So it seems to me that construction may matter more than the dielectric itself, at least as far as DF is concerned.
- keantoken
- keantoken
Hmm..must try those Vaseline caps one day!
It's puzzling though, what the bonding methods for mica dielectrics there actually refer to. Perhaps we can assume it ain't standard silvered mica type you're looking at there.
Actually. it should be realised that these will be lab. figures for a standard slab or volume of the material, not in cap form with all the associated series and parallel reactances, It looks at the substance in isolation.
It's puzzling though, what the bonding methods for mica dielectrics there actually refer to. Perhaps we can assume it ain't standard silvered mica type you're looking at there.
Actually. it should be realised that these will be lab. figures for a standard slab or volume of the material, not in cap form with all the associated series and parallel reactances, It looks at the substance in isolation.
Last edited:
Should be proportionnal to Dissipation factor , wich is given..
Or did i miss something.?...
This is untrue and misleading, polistyrene has much lower DA which is a big advantage when using pole zero compensation for vas circuits to optimise phase margins and distortion. DA causes time delay errors, papers have been shown numerous times on this forum about this subject. Even good quality high voltage COG caps outperform silver mica. Sonically polistyrene sound better or on par with the other cap types that outperform silver micas which are COG, polypropelene and teflon.
VAS miller compensation improves stability while decreasing OLG, lead compensation (I'm pretty sure that's correct) improves stability (or not) while increasing OLG. For improving stability, lead compensation tends to have a limited range of values, while VAS miller compensation can be applied to give more or less stability (although it is not always enough either).
I think lead compensation's more significant effect is to swamp out the feedback terminal's parasitics, which speeds up the amp and lowers parasitic-related distortions. So I think there are other good reasons for lead compensation than just stability.
To be clear, I am not talking about lead compensation where the compensation is taken from that VAS output. I am talking about when the main feedback resistor has the phase lead cap in parallel. In the former case my argument still applies, though you're trading feedback parasitic distortions for output stage distortions.
- keantoken
I think lead compensation's more significant effect is to swamp out the feedback terminal's parasitics, which speeds up the amp and lowers parasitic-related distortions. So I think there are other good reasons for lead compensation than just stability.
To be clear, I am not talking about lead compensation where the compensation is taken from that VAS output. I am talking about when the main feedback resistor has the phase lead cap in parallel. In the former case my argument still applies, though you're trading feedback parasitic distortions for output stage distortions.
- keantoken
Last edited:
Sonnya,
I have pretty much finalised my personal design which is slightly different to Hugh's because I wanted to run lower volt rails and a slightly simpler circuit.
I have also made my own PCBs for this circuit, but they are not to compete with Hugh's and are not for sale.
I got bored and messed around with a BJT cascode version but I didn't take it far.
Lately I have been messing around with output chokes with good results.
That's about whe things are at! I will post a new schematic soon.
I have pretty much finalised my personal design which is slightly different to Hugh's because I wanted to run lower volt rails and a slightly simpler circuit.
I have also made my own PCBs for this circuit, but they are not to compete with Hugh's and are not for sale.
I got bored and messed around with a BJT cascode version but I didn't take it far.
Lately I have been messing around with output chokes with good results.
That's about whe things are at! I will post a new schematic soon.
I repeatedly refer Members to Cherry's writing on the Thiele Network.
Thiele specified two types and it turned out these were the limiting conditions at either end of a continuum.
Cherry explains that continuum and gives a method for deriving typical component values that appear to work.
I took Cherry and realised that by combining the two limiting Thiele types into a Pi filter and using Cherry came up with values that also appear to work. I have posted this spreadsheet at least three times in the last 7years.
First Zobel 10r + 47nF
Inductor 0.5uH to 1uH // 2r0 to 5r0
Second Zobel (across the amplifier's speaker outlet terminals) , 68nF + remaining resistance to make up the R of L//R to the 10R value to match the first Zobel. Say 8r0 to add to 2r0, or 5r0 to add to 5r0, etc. The amp sees ~5r0+120nF as extreme HF load and the three resistors damp the reactive components so that they do not upset the amplifier, no matter what speaker or speaker leads are connected/not connected.
These component values are not cast in stone, all six values can be changed and even taken to their limits, eg. open circuit or short circuit (zero ohms) and still make up the Pi filter.
But, read Cherry.
Thiele specified two types and it turned out these were the limiting conditions at either end of a continuum.
Cherry explains that continuum and gives a method for deriving typical component values that appear to work.
I took Cherry and realised that by combining the two limiting Thiele types into a Pi filter and using Cherry came up with values that also appear to work. I have posted this spreadsheet at least three times in the last 7years.
First Zobel 10r + 47nF
Inductor 0.5uH to 1uH // 2r0 to 5r0
Second Zobel (across the amplifier's speaker outlet terminals) , 68nF + remaining resistance to make up the R of L//R to the 10R value to match the first Zobel. Say 8r0 to add to 2r0, or 5r0 to add to 5r0, etc. The amp sees ~5r0+120nF as extreme HF load and the three resistors damp the reactive components so that they do not upset the amplifier, no matter what speaker or speaker leads are connected/not connected.
These component values are not cast in stone, all six values can be changed and even taken to their limits, eg. open circuit or short circuit (zero ohms) and still make up the Pi filter.
But, read Cherry.
Last edited:
swordfish.....
pls be careful....the added coil in the output is a trick often employed with poor domes to make them perform a little better...especially hard membrane types seem to benefit from added inductance...while your system may sound better....it could just be adaption to your system....and not because the amplifier is improved...
pls be careful....the added coil in the output is a trick often employed with poor domes to make them perform a little better...especially hard membrane types seem to benefit from added inductance...while your system may sound better....it could just be adaption to your system....and not because the amplifier is improved...
Those were my thoughts to when adding a series resistor to my speakers. I removed it later and used the EQ in my media player. Though I suspect I would prefer a passive EQ to use at the input of my amps.
I've been thinking about getting some of the ribbon tweeters here to play with:
Speaker Stuff
I've heard tweeters do better with a high-impedance source, which may explain the inductor trick.
- keantoken
I've been thinking about getting some of the ribbon tweeters here to play with:
Speaker Stuff
I've heard tweeters do better with a high-impedance source, which may explain the inductor trick.
- keantoken
Homemodder,
This implies I am endeavouring to deceive people.
Nothing could be further from the truth.
It is MY experience over years in this game with all types of caps.
I offer it here without obfuscation or hype.
Have I been rude to you? If so, perhaps your comment is justified, though not in a public forum. Otherwise, I would like a retraction.
Hugh
This implies I am endeavouring to deceive people.
Nothing could be further from the truth.
It is MY experience over years in this game with all types of caps.
I offer it here without obfuscation or hype.
Have I been rude to you? If so, perhaps your comment is justified, though not in a public forum. Otherwise, I would like a retraction.
Hugh
Miib,
I actually think this might be the case with my speakers. As I have said repeatedly, my speakers are very bland sounding and I strive to liven them up a bit.
Everyone should be taking my opinions as mine only and try their own experiments. I post my results but can not stress enough that personal tastes and your own listening conditions may be different!
SWF,
The connection between measured specs and the 'sound' of a system, whatever that may be, is very difficult to pin down. You cannot argue with measurements, no question, but it doesn't mean that they describe or define the sonics. I am heartily sick of the prescriptive drivel from the math 'n measure guys who think that the only mark of excellence of an amp is the THD; things like distortion spectrum, cap choices, operating points, do make a difference and I for one always read your findings with interest and fascination as your experiences so often mirror mine.
I don't pretend to be right any more than you do, but I still think subjective impressions are very important.
Hugh
The connection between measured specs and the 'sound' of a system, whatever that may be, is very difficult to pin down. You cannot argue with measurements, no question, but it doesn't mean that they describe or define the sonics. I am heartily sick of the prescriptive drivel from the math 'n measure guys who think that the only mark of excellence of an amp is the THD; things like distortion spectrum, cap choices, operating points, do make a difference and I for one always read your findings with interest and fascination as your experiences so often mirror mine.
I don't pretend to be right any more than you do, but I still think subjective impressions are very important.
Hugh
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.