Three different "80 hy 25ma" plate chokes

Design and use of Chokes for PS, anode and filament use.
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Three different "80 hy 25ma" plate chokes

Post by dave slagle »

The tube to be loaded is an RS241 in various positions attached below are the L vs. I graphs for three different desings on different core materials. The graphs are pretty general and widespread at this point due to the range, but can be narrowed down in the future. The basic identification is by core material and DCR.

the first step is the selection of the initial gap for shipping and the first round of comparisons.

it will take some visual interpolation to read between the lines, but reasonable guesses can be made for any intermediate gap. Assuming the 80hy requirement, I'll suggest the initial gaps in the individual graphs and let me know where to do it and i'll add the gapped inductance as sent lines to the graphs with the closest over and under lines.
Attachments
updated with chosen gap
updated with chosen gap
49%.gif (12.27 KiB) Viewed 23348 times
80%.gif
80%.gif (12.1 KiB) Viewed 23348 times
amorph.gif
amorph.gif (11.88 KiB) Viewed 23348 times
Last edited by dave slagle on Wed Feb 09, 2005 4:50 pm, edited 1 time in total.
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

Post by Johnny »

Hi Dave,

Thanks for all the great work!!! Let's go with these values to start:

80% Nickel 0.024

49% Nickel 0.024

Amorphous 0.012


Are these values for both gaps added or just one?

Can't wait to try them out.

Johnny
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

i updated the images above to show what they are gapped at. I gapped the 80% nickel at .025 so there will be slightly less inductance than the .024 line on the graph.

the gap on the graph is double the actual spacer, so the 49% nickel chokes have a .025 gap which is 1 piece of "black" shimstock which measures .0125 inches.

i also attached the excel graphs for each so you can edit and print them out for your needs.

they are all set to ship, i just need to do some frequency sweeps on the 1:1 bifilar outputs.
Attachments
49%.xls
(16 KiB) Downloaded 424 times
80%.xls
(16 KiB) Downloaded 445 times
amorph.xls
(16 KiB) Downloaded 469 times
the 49 and 80% nickel chokes have a black shim for .025 and the amorph has a double green for .012
the 49 and 80% nickel chokes have a black shim for .025 and the amorph has a double green for .012
uscolor2.jpg (33.03 KiB) Viewed 23407 times
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

Post by Johnny »

Hi Dave,

Cool, looking forward to getting them. Thanks,

Johnny
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

Listened to the Chokes

Post by Johnny »

Hi Dave,

Grant Gassman, Steve Kaufman, and I got together at my place last Saturday to listen to the chokes you sent me. We used them in the driver stage of my amp. The choke loads a Telefunken RS241. The plate of the RS241 drives a second order filter ( a series cap and a parallel autoformer.) The 6db down point is 237Hz.
We first listened to the chokes with the gap they had when you sent them to me. Amorphous core, 0.012", 80% nickel, 0.025", and 49% nickel, 0.025". After comparing all of these to the amorphous Tribute tranformers used as chokes, I liked the amorphous Tribute transformers best by a reasonable margin. Next, I liked, in this order, the amorphous chokes, the 80% nickel chokes and the 49% nickel chokes.
Next we adjusted the gap in the amorphous chokes. We tried a range from 0.012" to 0.10". In the end, a gap of 0.06" gap sounded better than the other gaps we tried with the amorphous chokes by a wide margin. We definitely hit a sweet spot. The amorphous core chokes also sounded better than the Tributes by a significant margin, wasn't close.
Next we adjusted the gap in the 80% nickel chokes. We tried a range of 0.025" to 0.10". In this case, a gap of 0.05" gap sounded better than the other gaps we tried with the 80% nickel chokes by a wide margin. Again, it sounded like we "tuned" to a sweet spot. The 80% nickel chokes were also signifcantly better the Tribute amorphous core transformer.
Next we adjusted the gap in the 49% nickel chokes. We tried a range of 0.025" to 0.10". In this case, a gap of 0.075" sounded better than the other gaps we tried with the 49% nickel chokes by a wide margin. Again, it sounded like we "tuned" to a sweet spot. The 49% nickel chokes were also signifcantly better the Tribute amorphous core transformers.
Finally, we compared all three chokes after each was tuned to a sweet spot. I liked the 49% nickel chokes best followed by the 80% nickel chokes. The 49% nickel chokes seemed to capture more of the sound field than the other two. The reverb in a concert hall was more distinct and the strings sounded more natural. They also had more drive and seemed to handle transients better than the other two. The amorphous core chokes seemed a little muted to my ear.
I thought it was interesting that the worst sounding choke became the best sounding choke after adjusting the gap.
I know the gaps we picked are wider than you would expect Dave. I believe it's an "EQ" sort of thing. Since these chokes were used in an amp that only provides for frequencies above 230Hz, they don't need as much inductance. A lower inductance can be used to get a sound I like. The reason for the better sound might be that an agreeable resonant point was created by varying the inductance.

John
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Re: Listened to the Chokes

Post by dave slagle »

Johnny wrote:Hi Dave,

Grant Gassman, Steve Kaufman, and I got together at my place last Saturday to listen to the chokes you sent me. We used them in the driver stage of my amp. The choke loads a Telefunken RS241. The plate of the RS241 drives a second order filter ( a series cap and a parallel autoformer.) The 6db down point is 237Hz.
so it is essentially a parafeed coupled stage with no external load on the output of the transformer. (just want to get the details clear)
I liked the amorphous Tribute transformers best by a reasonable margin. Next, I liked, in this order, the amorphous chokes, the 80% nickel chokes and the 49% nickel chokes.
interesting order... what was the inductance and DCR of the tribute?
Next we adjusted the gap in the amorphous chokes. We tried a range from 0.012" to 0.10". In the end, a gap of 0.06" gap sounded better than the other gaps we tried with the amorphous chokes by a wide margin.
to be clear, was the 0.06" gap a single orange (0.03") shim?
We definitely hit a sweet spot. The amorphous core chokes also sounded better than the Tributes by a significant margin, wasn't close.
i suspect it has more to do with as you say a tweeky sweet spot than anything else. Pieter is a good friend and his stuff is top notch and he is welcome to all of the info you guys give... hopefully we can help everyone make better decisions.
Next we adjusted the gap in the 80% nickel chokes. We tried a range of 0.025" to 0.10". In this case, a gap of 0.05" gap sounded better than the other gaps we tried with the 80% nickel chokes by a wide margin. Again, it sounded like we "tuned" to a sweet spot.
how drastic is the sweet spot? i must admit, while i am thankful for your efforts, i am terrified by your results.
Finally, we compared all three chokes after each was tuned to a sweet spot. I liked the 49% nickel chokes best followed by the 80% nickel chokes. The 49% nickel chokes seemed to capture more of the sound field than the other two. The reverb in a concert hall was more distinct and the strings sounded more natural. They also had more drive and seemed to handle transients better than the other two. The amorphous core chokes seemed a little muted to my ear.
both the 49% and the amorphous have the same specs wrt DCR and operating flux and at the gaps and current you speak of they are both linear as hell. the 80% nickel has a higher DCR and gets run a bit harder.

in looking at the L vs. I plots both nickels show nice linearity and abrupt saturation, while the amorphous seems to just gradually coast into saturation. (just something i have observed)
I thought it was interesting that the worst sounding choke became the best sounding choke after adjusting the gap.
great, bizzaro magnetics... i hate you guys :-)
I know the gaps we picked are wider than you would expect Dave.
which is why i hate picking them... if i actually chose gaps that large i'd be drowned as a witch.
I believe it's an "EQ" sort of thing. Since these chokes were used in an amp that only provides for frequencies above 230Hz, they don't need as much inductance. A lower inductance can be used to get a sound I like. The reason for the better sound might be that an agreeable resonant point was created by varying the inductance.
agreed... you are still dealing with a very reactive circuit (LCL) and modifying any one of the elements can tune it. rather than swapping caps, you may effectively be doing the same thing by changing your first (plate) L. I do still believe that reducing the variability of any one of the elements in the LCL filter is a good thing and hope thats partially responsible for what you are hearing. to be clear, i trust 100% what you say you heard... now all we need to do is figure out how to make the results more repeatable.

thanks for the info!!! now we just need to coax a few number crunchers into the mix to help us connect the dots... i can draw a few fuzzy lines to connect the dots, but in general most people don't trust the word of someone who draws in crayon.

just for kicks, you should try the 1:1 bifilars as plate chokes and see how they sound with a few different gaps in this same situation... try them as IT's too, maybe getting rid of the cap will make everythng else moot.

dave
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

Post by Johnny »

so it is essentially a parafeed coupled stage with no external load on the output of the transformer. (just want to get the details clear)
Yes, that's right.

interesting order... what was the inductance and DCR of the tribute?
My amp is covered with jumper cables. I would have to disconnect them all and turn the amp over to get the inductance value of the Tribute. When I take the amp over to Steve's this Saturday, I'll get the inductance value and measure the drc.


to be clear, was the 0.06" gap a single orange (0.03") shim?
I used two shims, a white (0.025") and a blue (0.005").




how drastic is the sweet spot? i must admit, while i am thankful for your efforts, i am terrified by your results.
A difference of 0.005" was cleary audible. I know it doesn't make much sense. I'm glad Steve and Grant were there to hear it as well. Both said they would give their impressions as well.

in looking at the L vs. I plots both nickels show nice linearity and abrupt saturation, while the amorphous seems to just gradually coast into saturation. (just something i have observed)
With the original spacings ( 0.025"), the nickel chokes did sound harder than the amorphous chokes. Even after finding the sweet spot for all three, the amorphous chokes were smoother. I wonder if the nickel chokes with the narrow spacings were saturating during peaks in the music?
just for kicks, you should try the 1:1 bifilars as plate chokes and see how they sound with a few different gaps in this same situation... try them as IT's too, maybe getting rid of the cap will make everythng else moot
Should I put a resistor between one of the bifilar wires, connect them in parallel, or connect them in series?


Johnny
Grant
Posts: 7
Joined: Wed Feb 02, 2005 6:25 pm

Post by Grant »

The experiments were certainly very eye-opening to me. The way that John described it is not in any way over-dramatized. The differences among the small gap variations was VERY noticable and we all could tell when the sweet spot was reached without any head-scratching or debate - it was stark!

My personal preference was for the Amorphous Slagle chokes, as my taste runs more to the romantic than John's. They sounded more fleshed-out in the lower mids to me.

Since we were gapping these chokes to the extent that the inductance became far lower than the original spec, I am left wondering how things would've sounded with a choke wound for less inductance (for the same initial gap), since fewer windings would produce less capacitive losses in the high end.

Dave: any idea what the high-frequency roll-off point would be with the above configured chokes, assuming tube Rp of 5k or so? If it's over 100k already (-3dB), then probably not going to improve with less capacitance, but it was a thought I had...

John - It won't be easy to measure the inductance of the Tributes without a bridge...

Thanks for the opportunity to play with all the iron - very interesting! Looking forward to this Sat. when we should have a chance to play with the 1:1 trannies substituted for a cap in the parafeed opt stage!
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

Post by Johnny »

Hi Guys,
A difference of 0.005" was cleary audible. I know it doesn't make much sense. I'm glad Steve and Grant were there to hear it as well. Both said they would give their impressions as well.
I made a mistake here, the difference should be 0.010" because there was a difference of 0.005" in each gap for a total difference of 0.010"

John
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

With the original spacings ( 0.025"), the nickel chokes did sound harder than the amorphous chokes. Even after finding the sweet spot for all three, the amorphous chokes were smoother. I wonder if the nickel chokes with the narrow spacings were saturating during peaks in the music?
i really doubt it... on my way to work tomorrow i'll plot the inductance vs. gap, ac and dc flux behaviors. this will all be by the math, but it lines up really close to what i measure. what is the max AC voltage swing they see???

Should I put a resistor between one of the bifilar wires, connect them in parallel, or connect them in series?
yes :-) seriously i would parallel them, and maybe try a single wire with the second unloaded to see if there is a difference. just for kicks you can try them in series, but i would expect some wacky capacitance issues.

dave
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Grant wrote: My personal preference was for the Amorphous Slagle chokes, as my taste runs more to the romantic than John's. They sounded more fleshed-out in the lower mids to me.
really?? i have always found nickel to be more romantic... hmmm...
, since fewer windings would produce less capacitive losses in the high end.
how about if i send you guys a pair of bobbins for the 49% nickel with 1/2 the turns??? seems like a worthwhile swap.
Dave: any idea what the high-frequency roll-off point would be with the above configured chokes, assuming tube Rp of 5k or so? If it's over 100k already (-3dB), then probably not going to improve with less capacitance, but it was a thought I had...
actually i have no idea... i would expect 100K+ for these particular ones.
Thanks for the opportunity to play with all the iron - very interesting! Looking forward to this Sat. when we should have a chance to play with the 1:1 trannies substituted for a cap in the parafeed opt stage!
so am I. also try it as a straight plate choke and play with the gap. lastly, when you do try them as outputs and dismiss with the cap, be sure to ground the secondary for safety reasons. I have had 4000V between the windings without issue, and ran a gm-70 at 800V for an hour, but you can never be too safe.

dave
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

Post by Johnny »

how about if i send you guys a pair of bobbins for the 49% nickel with 1/2 the turns??? seems like a worthwhile swap.

Sounds great!!!! Let me know when you send them. Thanks,

John
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

hey guys.

did some calculations on what the gaps you chose are giving you.

Amorphous .06 gap.
14.3hy
1440 gauss Bdc @ 25ma
2200 gauss Bac 50V@20hz

49% nickel .075 gap
19.6hy
1319 gauss Bdc @ 25ma
1500 gauss Bac 50V@20hz

80% nickel .05 gap
33.8hy
2472 gauss Bdc @ 25ma
1633 gauss Bac 50V@20hz

the range of final inductances suggest that maybe it wan't the tuning of the filter. I chose 25ma and 50V @ 20hz as arbitrary points to show that you are well away from saturation (14KG for the 49% and amorph and 7KG for the 80%.

What was the DC current and AC voltage for the test? also the lowest frequency where full signal voltage was present?

i have the bobbins with 1/2 the turns done... i'll try to get them out tomorrow.

dave
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

Post by Johnny »

Hi Dave,

Your calculations are very interesting. I think you're right that it may not be a tuning issue. So many things in audio have no solid electrical explaination. For example, why do some tubes sound better than others (maybe harmonic content?), why do tube amps sound better than solid-state amps (most solid-state amps have negative feedback that can cause intermodulation distortion?), why some caps sound better than other caps (I have no clue), why do triodes sound better than pentodes (maybe better impedance matching with amp-speaker interface?), why do directly heated triodes sound better than indirectly heated triodes (a complete mystery), why does a digitally recorded performance played back on LPs sound better than the same digitally recorded performance played back digitally. I could go on and on. In the end, I think the area of psycho-acoustics has the most promise for explaining what we hear and prefer.
What was the DC current and AC voltage for the test? also the lowest frequency where full signal voltage was present?
The DC current was 15ma. I'm not sure what the AC voltage swing was but it was no where near 50V. The 2nd order filter (parafeed configuration) that follows the choke has a 6db down point of 237hz. Since the choke is part of the driver stage that preceeds the 2nd order filter, the choke could still "see" 20Hz. Since it looks like your calculations are conservative, I don't think the cores are saturating.
Thanks for the calculations. It makes this a lot of fun!


John
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Your calculations are very interesting. I think you're right that it may not be a tuning issue. So many things in audio have no solid electrical explaination.
Luckily we don't need an exact explanation but to see a pattern emerge. Low DCR chokes tend to sound better to me, choke input filters, DHT's and removing caps and resistors all make a positive impact on the sound. I know this from listening experience and do not need a proof. This doesn't mean I won't continue to think about it to attempt to connect the dots.

A perfect example of this is the low DCR PS choke and your gapping results. The whole DCR as a factor of merit has gotten a lot of play generally dividing the "numbers" people from the "ears" people. Personally I think the numbers camp has this one all wrong. The ears people have come to this conclusion by listening and making the observation that "low DCR" = "better sound" the numbers people cannot stomach such nonsense so the do everything they can to prove how designing for low DCR can give you god awful results proving the ears cap wrong. Talk about twisted logic! when you look at the chokes the low DCR people are preferring, 9 out of 10 times they follow the basic rules of choke design. There is a cookbook if you will that gives the optimal choices for a particular design. Have you ever noticed that a 10hy 200ma choke always seems to have a DCR of 150 ohms independent of who made it? Why?? Its a design straight out of a cookbook. Now say you want a 10hy choke with a lower DCR, when you go searching through all of the cookbook designs out there, you end up finding a 10hy 500ma choke that meets your desired DCR spec and accept the added size, weight and unused DC current potential to meet your DCR spec. the important thing to note is along with the DCR a lot of other factors have changed. The AC flux for your ripple, the DC flux from a fixed current (80ma??) and the size of the airgap will all be different, still the easy thing to identify is the DCR.

The important thing to note is all of these chokes are designed to a certain set of compromises and lowering DCR always gets you a bigger higher current design as long as you don't cheat, and trust me getting a transformer designer to cheat is like pulling teeth :-) So where the hell am I going with this??? If we assume none of the golden rules are broken, a low DCR choke will operate at a lower AC and DC flux and have a *larger airgap* than its higher DCR brother. Now we have you crazy guys breaking some rules and using a larger airgap than anyone would ever dream of and I suddenly saw some dots that might need connecting.

I have often thought that selection of an airgap should be the choice of compromise between inductance and linearity. Increasing the gap increases linearity, but at some point you will go to far and the lack of inductance will be your main concern. You guys have seemed to show this pattern, but the three different inductances of the "tweeked" chokes doesn't help :-(

Since this is a parafeed circuit, the DCR does come into play, and its interesting to note that in each case the preferred gap had some relationship to the DCR. In each case as the inductance went up, so did the DCR of the coil in question. Starting with the amorphous and 14hy 900 ohm, we go to nickel 19hy 1000 ohm and 80% nickel 33hy 1700 ohm. The 49% showed a minor increase in both DCR and L over the amorphous, and the 80% is roughly double... Just an observation. I also want to be clear that 50% one way or the other is not significant since there are so many places for errors in the measurements and calculations.

Oh boy... Now back to where I was initially trying ot head before I was sidetracked.

If the tuning of the filter (and crossover) isn;t what you are hearing, I had to dig into my bag of dots so see what else we might be able to connect. The best one I could latch onto is this is the old single quadrant of operation.

For a given AC flux, as you increase the gap size your DC flux will get smaller as the AC flux remains the same. So with a very small gap your Dc flux is much greater than your AC and the only concern is hitting the knee of the BH loop. As you increase the gap at some point the Bdc and the Bac will be equal, and this represents the limit of single quadrant operation. A further increase of the gap will lower the Bdc and allow the AC flux to operate in all 4 quadrants of the graph. Ie the AC will pass the dreaded "zero crossing" point. I ran the numbers and at 15ma the point where Bdc and Bac are:

Amorph: 220V@237hz
49%: 300V@237hz
80%: 525V@237hz

Oh well... It was an idea :-(

The DC current was 15ma. I'm not sure what the AC voltage swing was but it was no where near 50V. The 2nd order filter (parafeed configuration) that follows the choke has a 6db down point of 237hz. Since the choke is part of the driver stage that preceeds the 2nd order filter, the choke could still "see" 20Hz. Since it looks like your calculations are conservative, I don't think the cores are saturating.
Well, the formula for AC flux is linear, so going from 237hz to 23.7 hz just requires a move of a decimal point. Then we are into numbers that may actually be relevant. Also if there are any bass peaks from the resonant circuit, they can give you much larger ac flux values than simplemath would predict. I do agree that we are far away from saturation, but we still may be hovering right around where Bdc = Bac and or the zero crossing point.

dave
Steve Kaufman
Posts: 35
Joined: Wed Jan 26, 2005 6:00 am
Location: Fort Collins, Colorado

Post by Steve Kaufman »

Hi Dave

Grant, John, and I had a marathon session at my place yesterday. First, we substituted an inexpensive German-made preamp which I acquired in October. It is a tiny unit with a separate power supply and utilizes subminiature tubes. It sounded much better than my regular preamp, the construction of which you were aware of. The tubes (6021,6111) were apparently developed for military purposes and have very high bandwith and low microphonics. So much for all the theory that went into the development of my recently completed preamp! Needless to say, we stuck with the German preamp for our session.

John set up his amp and we tested the "80Hy chokes." Initially we plugged them in with the gaps as they had been optimized in John's system. The same "swet spots" applied in my system and the relative sonic characteristics between the different materials also carried over. Grant and I liked the amorphous chokes the most and John preferred the slightly more detailed sound of the 49% nickel chokes. When we narrowed the gaps on a pair of chokes to increase the inductance the high end suffered.

The bifilar 1:1s we used to eliminate the parafeed cap in the output stage severely compromised the sound, especially in the midrange and upper registers. However, the most surprising finding based on earlier listening, was that your amorphous output transformers used with the gaps you had specified were substantially better than the Ulbrichs used in parafeed. The cap is gone and the amps just got lighter!

For the purpose of building a pair of amps for my system some decisions have been made, namely amorphous output transformers and driver stage chokes. However, there is a major difference between John's application of these amps (he only uses them down to 200Hz) and mine (my amps run full range). If I choose to run my system with a single full range amp then somehow the inductances need to be increased without jeapordizing the magic that the larger air gaps impart to the sound of the midrange and top end. In theory an inductance or 18Hy leads to a rolloff at about 51Hz. I could run it this way and boost my subs which kick in at 100Hz but the augmentation would not be linear (the higher frequencies should be boosted more than the more rolled off lower ones). We probably need to measure the output from John's amp and see how low the amp goes and there may be some effect in my listening room that causes us to seemingly hear below 50 Hz. In any case I would welcome your opinion on how to increase the low end extension without compromising the beauty these trials have produced. We are all grateful for your expertise and trannies and chokes.

Thanks

Steve
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Steve Kaufman wrote: First, we substituted an inexpensive German-made preamp which I acquired in October. It is a tiny unit with a separate power supply and utilizes subminiature tubes. It sounded much better than my regular preamp, the construction of which you were aware of. The tubes (6021,6111) were apparently developed for military purposes and have very high bandwith and low microphonics. So much for all the theory that went into the development of my recently completed preamp! Needless to say, we stuck with the German preamp for our session.
interesting... lets not throw the baby out with the bathwater though... otherwise we could all be listening to PP pentode amps... or worse yet solid state. The kind of wakeup call you seem to have had is a bit tough to take, it really makes you rethink a few things, but it shouldn't make you jump ship from everything you have experienced up to that point. you now have a goal to better, i wouldn't dump all of your tele DHT's for some milspec IDT's (although they are for the most part free in comparison)
John set up his amp and we tested the "80Hy chokes." Initially we plugged them in with the gaps as they had been optimized in John's system. The same "swet spots" applied in my system and the relative sonic characteristics between the different materials also carried over.
that is reassuring.
When we narrowed the gaps on a pair of chokes to increase the inductance the high end suffered.
somehow i am both comforted and terrified by this statement. I trust your report, and can easily attach some theory, but the theory scares the hell out of me when it comes to reality (which you are alluding to below)
The bifilar 1:1s we used to eliminate the parafeed cap in the output stage severely compromised the sound, especially in the midrange and upper registers.
bummer... did you try them as plate chokes? did you play with the gap?
However, the most surprising finding based on earlier listening, was that your amorphous output transformers used with the gaps you had specified were substantially better than the Ulbrichs used in parafeed.
I'm not exactly clear what this means. I didn't realize you had compared direct feed to parafeed. could you referesh us with that experience?
The cap is gone and the amps just got lighter!
because you also eliminated the plate choke?
If I choose to run my system with a single full range amp then somehow the inductances need to be increased without jeapordizing the magic that the larger air gaps impart to the sound of the midrange and top end.
you can always go for more turns on a larger core, and amorphous makes this easier, but at some point, too much wire on too big of a core will compromise the top end. I don't believe anyone has evere explored where that point is. the common wisdom is use a small gap to get as much inductance as possible so you don't end up compromising the top end. Then you guys come along and say a larger gap makes the top end sound better. If it's any comfort, people have had similar results with the autoformer volume controls noting that a larger gap makes things much better.
In theory an inductance or 18Hy leads to a rolloff at about 51Hz. I could run it this way and boost my subs which kick in at 100Hz but the augmentation would not be linear (the higher frequencies should be boosted more than the more rolled off lower ones).
on the surface it is a simple choice of compromises. the problem becomes identifying the important compromise on an individual case by case basis.. the issues caused by too small of an airgap seem to be a priority based on your limited tests, but this shouldn't become the only driving force in your design.
In any case I would welcome your opinion on how to increase the low end extension without compromising the beauty these trials have produced.
if we keep the idea of a large airgap as a priority, you have two options to increase the inductance. a bigger core and/or more turns. a first place to look for some insight would be trying the bigger core approach. If you parallel the bifilar 1:1 windings you have a fair example of the bigger core approach. Just for kicks you could also try to series the windings and increase the gap more. My gut feel is this will cause more harm than good, but what do I know?

dave
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

here are the measurements of a bobbin for the 49% nickel cores with fewer turns of thicker wire. I'll get them sent out in the next day or so.

enjoy!
Attachments
the DCR is about 300 ohms or so.
the DCR is about 300 ohms or so.
pasetto.gif (8.59 KiB) Viewed 22355 times
pasetto.xls
this is the excel graph of all of the gaps i measured and should be tweekable in excel.
(the scales, size, etc)
(15 KiB) Downloaded 382 times
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

Post by Johnny »

Thanks a bunch Dave! I'm excited to try these. What I want to do is keep the inductance for the bobbin with half the turns about the same as the inductance with more turns. The bobbin with half the turns should have a narrower gap than the other one. I think this comparison should give us some good information concerning how the gap independently affects the sound. It's going to be fun.

John
Steve Kaufman
Posts: 35
Joined: Wed Jan 26, 2005 6:00 am
Location: Fort Collins, Colorado

80 Hy chokes

Post by Steve Kaufman »

Hi Dave

John was over yesterday. He enlarged the gaps in your 80Hy amorphous chokes in the line stage of my preamp from 0.01" to 0.06". The effect was exactly the same as in the amp trials we have done recently. The sound "opened up" considerably with much more natural sounding vocals, woodwinds, and strings and no loss of dynamics. We had to add a little volume to the subwoofers, essentially below 100Hz but bass was no issue.

Steve
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Re: 80 Hy chokes

Post by dave slagle »

Steve Kaufman wrote: The sound "opened up" considerably with much more natural sounding vocals, woodwinds, and strings and no loss of dynamics. We had to add a little volume to the subwoofers, essentially below 100Hz but bass was no issue.
Its amazing what such a simple listening impresison can suggest. The idea that a pattern is emerging is quite exciting and starts to give credibility to all your tests thus far.

The most interesting thing to me was that you got to the point where you lost bass, but were able to correct for it by adjusting a sub. This seemingly simple statement holds such deep ramifications that I am scared to voice them for fear of being struck dead on the spot :-)

If you got to the point where you lost bass i assume your inductor started (or helped) the low pass before the frequency it was set for. If your horns go full range and the subbs just "fill in" then your inductance rolled things off before the natural horn rolloff. either case opens up some interesting options for the future.

the piece of conventional wisdom that you are challenging is the use of a limited inductance plate load to act as a low pass. The whole battle cry of the large inductance crowd is if you don't have enough inductance massive distortion is sure to happen in the lowest registers and that will cause massive distortion in the higher registers.

I have never bought that position, and by that logic if you increased the gap so much that you lost bass, surely you would hear it in the upper frequencies. Well you did hear it :-) and your results happen to be the exact opposite of what some would predict.

Thanks!
sbench
Posts: 296
Joined: Fri Jan 28, 2005 3:45 pm

Post by sbench »

The most interesting thing to me was that you got to the point where you lost bass, but were able to correct for it by adjusting a sub. This seemingly simple statement holds such deep ramifications that I am scared to voice them for fear of being struck dead on the spot
Don't mean to hijack the thread, but I JUST COULD NOT keep my fingers from typing on this one.

Two comments, then I'll shut up.
1. Don't forget that as you lose bass response, you're also extending the HF response. Transformer BW tends to be constant octave width, so that if you lose an octave on the bottom end, you pick up (not quite) an octave on the top end.
2. My T-Lines are designed specifically with slightly rolled off Bass that is recovered with a <shudder> bass tone control. That's done absolutely on purpose. I've found over the years, it leads to tighter and more natural LF. Is that against common wisdom? Probably so. So what. It sounds good.


Sorry for jumping in. Back to lurk mode on this thread.

Steve
Steve Kaufman
Posts: 35
Joined: Wed Jan 26, 2005 6:00 am
Location: Fort Collins, Colorado

Post by Steve Kaufman »

Dave

My AvantGarde Trios do not run full range. The largest horns drop off at about 100Hz and the subs take care of the rest. The frequency and volume of each of the four subs is adjustable. We never heard the distortion at either end. Even with the reduced inductance the affected frequencies would be well below 100Hz. If the low frequency starts to drop off significantly at around 50 Hz, then boosting the frequencies between 20 and 50 to normal levels should boost the frequencies between 50 and 100 Hz to artificially elevated levels. The thing is, it sounds fine. Like most other listening rooms mine isn't perfect. Perhaps the qualities of my room smooth out this area. I played a large orchestral piece and a White Stripes album tonight and they sounded sharp as a tack.

Mr. Bent, thanks for your comments.

Steve
Steve Kaufman
Posts: 35
Joined: Wed Jan 26, 2005 6:00 am
Location: Fort Collins, Colorado

bob bob bobbin away

Post by Steve Kaufman »

Hi Dave

John received your bobbins for the output trannies. We listened to the original OPTs with optimized gap and then substituted the new bobbins with fewer windings and narrowed the gap to I believe 0.03". The latter provided a substantial improvement in high frequency and low frequency response as well as in openness and speed. Maintaining the inductance and lowering the DCR really helps!

Steve
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

Post by Johnny »

Hi Dave,

I agree with Steve, the 49% nickel chokes sounded better when we used the bobbin with half the windings and a narrower gap. As Steve said we compared the 49% nickel chokes with a gap of 0.075" and more windings to the bobbin with half the windings and a gap of 0.030". The bobbin with fewer windings was more open and at ease, without any harshness.
I was curious, so today I varied the gap on the bobbin with fewer windings. The sweet spot for me turned out to be 0.025". It had a little more "juice" than the 0.030" gap but otherwise pretty close. I tired narrower gaps, 0.010" and 0.020" but they sounded closed in compared to 0.025".
I think several things are going on here. First, since the DCR of the choke with half the windings was a third of the DCR of the choke with twice as many windings ( ~277 ohms to 921 ohms), the response could be quicker. I know you are a proponent of low DCR where posssible Dave. Second, after talking with Grant, I think fewer windings gives a higher frequency response because the intra-line capacitance is lower. Finally, I think a wider gap, up to a point, changes the distortion characteristic for the better.
If you can get away from using a lot of inductance, I believe the overall sound will be much better. I can do this in my rig because I bi-amp. I also think it is a great advantage to adjust the gaps in your rig for the best sound. Real interesting!

Johnny
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

I was curious, so today I varied the gap on the bobbin with fewer windings. The sweet spot for me turned out to be 0.025". It had a little more "juice" than the 0.030" gap but otherwise pretty close. I tired narrower gaps, 0.010" and 0.020" but they sounded closed in compared to 0.025".
Very cool! you seem to have a pattern that keeps repeating itself!
First, since the DCR of the choke with half the windings was a third of the DCR of the choke with twice as many windings ( ~277 ohms to 921 ohms), the response could be quicker. I know you are a proponent of low DCR where possible Dave.
Yes... Assuming all things are equal. But I think too many people take this to the extreme. The next logical test would be to go another step and cut the turns in half again. This will get you another drop in DCR and require a smaller airgap to get the "needed" 20hy of L... Are you game??
after talking with Grant, I think fewer windings gives a higher frequency response because the intra-line capacitance is lower.
Of course! its interesting to note that simply measuring FR of both will still net you bandwidth out past 100K! people always seem to forget what happens on the bottom end with the ellipse, also happens on the top end too as the load becomes capacitive, and this could easily explain why the greater top end extension of the fewer turn bobbin was audible since it would make the load at HF appear less capacitive.
Finally, I think a wider gap, up to a point, changes the distortion characteristic for the better.
I agree and think your experiments all fit nicely with what theory would predict.

There are dozens of variables to chose from when designing a choke and I suspect that every one of them will show you a "null" or point of optimal sound as you vary it from one extreme to the other.

Your first results came with everything fixed and the isolated variable was Inductance Vs. Linearity. Logic says the larger the gap the better because it will give you a more linear device. I believe that for a specific situation you can have too much inductance, and the price for this is simply a compromise on one of the other variables.

I am betting that your ideal gap for the first bobbin was simply the largest gap that still provided the required inductance for your specific situation and going beyond that gap just caused the lack of inductance ot be the major factor in the sound. What your next test with a lower DCR bobbin suggested was that there were other selections of compromise that were more important and could offer further improvement.

If you look at the graphs for the lower DCR bobbin, you are still in a very linear part of the curves and it would take 4X the DC current that you are running to get to the point where saturation begins. This makes me wonder how an 80-100 ohm bobbin with a .006 inch gap would sound. The goal here would be to find the point where we get to too small of a gap (where core non-linearity bites us) and if we find that point it should cement down at least part of the process and maybe a few simple ROT's can be devised for future attempts.

The choice simply becomes that of juggling DCR, Inductance and linearity. Optimizing for any two compromises the third. (I left capacitance out of this but it can roughly be included in DCR, since for a given core the DCR and capacitance are proportional. DCR is just an easier measurement to grasp. I think its fair to assume if DCR goes up, so does capacitance (half the turns of a thicker wire will have a lower DCR and less capacitance)
I also think it is a great advantage to adjust the gaps in your rig for the best sound. Real interesting!
Agreed... Now all we need is a few others to play with some gaps and see if we do indeed see a pattern emerge. The compromises are there and there is no getting around them, but that doesn't mean that we can't beat a few of them into submission and take out some dogma at the same time :-)
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

Low Turns Experiment

Post by Johnny »

Yes... Assuming all things are equal. But I think too many people take this to the extreme. The next logical test would be to go another step and cut the turns in half again. This will get you another drop in DCR and require a smaller airgap to get the "needed" 20hy of L... Are you game??
Yes I am game! This is a lot of fun. Send out a bobbin with the turns cut in half again when you get a chance. Thanks,

Johnny
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

Johnny

Post by Johnny »

Hi Dave,

Grant and I got a chance to try the driver bobbins with a 1/4 of the turns. The difference was huge!!! The bobbin with a quarter of the turns was fast and not a bit hard. There was a ton of air and reverb. We used a gap of 0.010" with the 1/4 turn bobbin. The half turn bobbin had a gap of 0.025". We didn't have enough time to try different gaps for the 1/4 gap bobbin. We were shocked by how good our first guess was. The half turn bobbin sounded dark and sluggish by comparison.

We also tried the output bobbin with half the turns, (25:1 ratio). Again, this change was significant. More open and nibble. The full turns version sounded blurred by comparison. The full turn bobbin used a 0.040" gap and the half turn bobbin used a 0.020" gap.

I'll try "tuning" the gap for the 1/4 turn bobbin and the 1/2 turn bobbin in the next couple of days.

Johnny
dave slagle
Posts: 2038
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Re: Johnny

Post by dave slagle »

Johnny wrote:Hi Dave,
Grant and I got a chance to try the driver bobbins with a 1/4 of the turns. The difference was huge!!!
interesting, the DCR is of course lower which could be part of the explanation, another thing i have been thinking about has to do with the AC flux levels involved.

as you reduce turns you increase the AC flux and end up using a larger part of the BH loop. most will suggest this causes more distortion, but since oyu guys have a large gap the core is pretty linear anyways.

many hae said that the barkhausen steps that make up the loop are really similar to digital, so maybe by increasing the flux you are increasing the number of steps for the small signals (increasing resolution)

this pretty much goes opposite of what everyone says, but for some time now, i have been thinking if you need a 2W transformer that you should design it so that 2W uses as much of the loop as possible, so the low level signals will also use a larger share of barkhausen jumps per cycle.

just tossing that out there, but suffice it to say that i do not agree with the logic that using a 20W transformer on a tube that can only deliver 2W is a good thing because it reduces distortion.
The bobbin with a quarter of the turns was fast and not a bit hard.


what i tend to find as i reduce DCR.
There was a ton of air and reverb.
this is what makes me think you might be in a "better resolving" part of the core.
We also tried the output bobbin with half the turns, (25:1 ratio). Again, this change was significant. More open and nibble. The full turns version sounded blurred by comparison. The full turn bobbin used a 0.040" gap and the half turn bobbin used a 0.020" gap.
cool... same results as the plate choke. At lease we see a pattern that crosses the material boundry (plate choke was nickel and the output amorphous)

the other thing that comes to mind is both the low turn choke and the OT have better HF response than their higher turn brothers, but that would take some measurements (in circuit) to determine.
I'll try "tuning" the gap for the 1/4 turn bobbin and the 1/2 turn bobbin in the next couple of days.
one other thought about the gap tuning, is you are also changing the resonant frequency. the C is fixed and as you adjust the L the Fres is moving. no idea what this means for the sound though its just another thing i would like to look into. i'm thinking if we put Fres at 2K we will have equal loads at 20hz and 20K (20hz inductive, 20K capacitive and at the Fres it is resistive) so as you increase the gap, you reduce low end bandwidth pushing the Fres up slowly making the capacitive load at 20K become more resistive. just thinking out loud.... sorry if it makes no sense.

in any event keep up the good work.


dave
Johnny
Posts: 62
Joined: Wed Jan 26, 2005 3:43 pm

Post by Johnny »

as you reduce turns you increase the AC flux and end up using a larger part of the BH loop. most will suggest this causes more distortion, but since oyu guys have a large gap the core is pretty linear anyways.

many hae said that the barkhausen steps that make up the loop are really similar to digital, so maybe by increasing the flux you are increasing the number of steps for the small signals (increasing resolution)
I think you're on to something. I've listened to 845 rigs with big Audio Note transformers. They have many great attributes but their low-level resolution isn't as good as what I'm hearing now. I orignally started with the Tarmuras because they sounded very good in low power appplications. They are limited to 10Watts I believe. The tube I'm using, 75TL, is capable of more power but I think it would be a mistake to use more power because the bigger transformers wouldn't be delicate.


John
Post Reply