@DaveyF: Totally understand the risks. I think the circuit is in many ways superior to the Classic 120s and VTXXXs I used to own, and provides more opportunities for asjustments, thats why my quest for this amp.
@HifiGeek1: Thx for the cautions. Guess I have been lucky. The only Hexfreds that have exploded on me was the IXYS bridges. Been lucky with fairchild 1200V stealth and ST 100V Schottkys. Knock on wood. The filament supply is good for 4 x 625mA (6FQ7), AC filament in parallel, and I am only changing 2 of those guys to 900mA. The 6922s use DC filaments (15VAC winding) and thats where I plan to use the 100V Schottkys. |
Anyone want to part with their D70? No tubes OK (even better). Itching to do a total rebuild project with it. |
Thanks for the opinions. Now to find an old beat up unit for an affordable price so I can do a complete rebuild. |
Found a D115 mkii. Can't wait to hear how it sounds.
Will definitely verify each test points voltage and go through the usual hexfred/schottky diode mods and coupling cap upgrades... then triode of course.
The circuit begs for a better driver tube than 6FQ7.
Will see if filament supply can support 6H30 (driver and power tubes have AC filament). |
@Hifigeek1: Yes indeed triode vs NON-Ultra-linear fixed bias tetrode mode. But same B+ @ output stage.
420V B+ is a bit low for my liking but with 4 tubes per side power should be more than enough.
I have every intention to convert that to triode at a later stage, after I am done with the caps and diodes.
I had the CL120 and CL150s before. CL120 at least has a tube driver stage (different plate voltage from D115 though). CL150 is just raw muscle. |
Oh well. So much for the exciting anticipation for the amp.
Seller allegedly got talked OUT of including the original outer box by the FedEx people and the unit was shipped in the inner box only with less than a few pages of newpaper in the box so go figure what happened. |
all measured OK. Thanks HifiGeek1 for the pointers.
Took me 2 hours to carefully pull all the wondercaps and Relcap coupling caps out, and another 2 hours almost to install k40y9s into the amp (size and metal casing does NOT help at all).
The foggy when low volume and glare at high volume problem is mostly gone.
I get a lot more natural and airy sound with strings.
More to do but at least this IMO is going absolutely the right direction. |
So the verdict for my unit is:
Faceplate is dinged up and dented. Supporting bar bent. I cant get a replacement from Audio Research as they are out of spares.
Seller graciously offered a partial refund which is very nice of them.
The 1A fuse popped several times and made me worried the screen regulator circuit might have hidden issues. 3 x 1A fuses and a trip to the local store with a tube tester I am fairly certain it was a bad 12AT7 that tests weak on the tester. Replaced it and I have powered it up 10 times since with no issues.
I am hoping that the fuse pops were just a problem with the tube.
Since the unit is of late vintage and confirmed to be a mkII I think the lytics are still OK.
I meticulously adjusted the diff amp balances and bias and now it sounds better than initially, and so far the voltages (esp screen regulator output I am cautious of) look stable.
Over sound is that it's really foggy and a bit slow. Crank it up and I notice glare and roughness.
Space is real tight for any cap mods. Auricaps will fit but how much better are they than the Rel + Wonder paralleled.
A friend has a pair of Reference 300 that's in the store room for at least 4 years now since the last power up. Maybe this little project, if successful, can get him to power them up and play again.
@HifiGeek1: What do you think of the recent frenzy regarding those whatever 50 years treasure funny black bottles and the whatever black sable 6550s? PM me if you want to avoid flaming :) |
All tubes are known good tubes from my collection. All voltages were checked and verified AOK.
I have no way of checking how good the lytics are cos I dont have a big cap meter and I dont want to pull out the lytics.
Foggy is expected of the Rel-cap/Wondercap combo, as I have over the years owned quite a number of Audio Research amps. No surprises there. With this amp, everything is like 95% there but not quite excellent. Nothing annoying, but nothing that leads me to say, wow this track plays good.
CL120
VT150
CA50
VT200
CL150
VT130
and I might have missed a few...
Most of them I ended up redoing the coupling caps with Infinicap-S, Dynamicap, Auricap, or Jensen (wish I had tried V-caps and worked on the all-tube models more before I sold each of them one after the other).
Problem right now is that space is at a premium and its very difficult to put in premium caps into this amp due to their size. |
Point well taken, esp with coupling caps where I often choose not to parallel a tiny 0.01 for the sake of coherence. Would rather invest more and get a better single cap.
For B+ bypass my experience is YMMV. I noticed that blackgate lytics will sound hashy and smear if bypassed with a small film whereas others may benefit.
the D-115 B+s are all lytic + Rel2.0uF + Wonder 0.01 with regulated B+ to the screen rel2.0uF + wonder 0.01.
I also notice that the B+ to the 6922 gain stages have nothing at regulator (TIP41) output at all (!).
I think I am ready to move on to triode mode. Just realized the screen grid supply is actually the 6550 pass element's output. This will be useless if I convert to triode, kind of a waste.
Wonder why they use the solid state buffered regulated output for the small signal B+ and not directly from tube reg. |
@DaveyF: No other offender. I am comparing it with my 6 other amps in the room.
Vac Phi 70 monos
Counterpoint NP220 Premium gold
Jolida 502 thats "re" modified from an already modified unit
just to name a few.
@hifigeek1: Thanks a lot for the suggestions.
I myself hate using NOS power tubes, I pity those who generic Sovteks and Svetlanas remarked as whatever the hell reissue and give them names like Black voodoo and blue snake oil. Chinese tubes even to this date are a complete joke. Just sad to see the feeding frenzy for those 50 year treasures. I myself use Svetlanas or Sovteks. They dont sound as good as the NOS stuff but with balancing and tweaking of the gain and driver stages I can live with the imperfections of run of the mill and reasonably priced power tubes.
Thank you very very much for the pointers. I scope the regulated B+s carefully. Thanks for the tip on measuring the 27k resistor. Been bitten before by a few resistors on my VT130 which measured wierd...disintegrated when I tapped them with a chopstick.
I am giving up on Hexfreds on this one, no way I can safely mount them unless I can find ones in smaller than T0220 package.
Thanks again. |
@Hifigeek1 you are talking about the 270k drop resistor R75 thats a protection bleeder across the main B+ filters right? I cant seem to find a 27k in the schematic. Is that a WW or a Carbon comp? |
Hmm I think I found it, goes from ground to pin4 of TL071, big 7W WW resistor. 26.4k and holding stable no change with tapping. |
I wouldn't put 8 kt120s into it. Consider how much more current the filaments will draw. 8 kt88s in tetrode mode should be able to power the hungriest of hogs. The apparent lack of power is from the stock caps. If you are not comfortable with DIY or non factory mods, ask Calvin at audio research if they offer recapping service for the coupling caps like they did i(infinicap and dynamicap) in the 90s. The relcap plus wonder cap combo is not conducive to good sound at all. |
@sbrown if you are just after the sound of the kt120 and not the marginal increase in power you can use just 4 tubes in your 115. That's how I sometimes run my old vt130 anyway. Make sure you are using one tube per phase per channel. Filament supply is parallel so you will never be stressing the unit with half the power tubes. |
Checked all the regulated B+s and they are quiet and stable. No more oscillations.
All diff amp balances measure AOK.
Gain is too high for my liking so in a sense it's indeed not quiet enough.
Maybe a triode mod will give me the extra benefit of lowering the gain.
The sonic signature is not lightweight and refined enough, lots of power though. I guess its what full pentode/tetrode mode is all about. |
D115... Becomes a tubefied CL60... |
Sovtek and Svetlana tubes are definitely good value for the money. I have owned no less than 8 Audio Research amps and I have only 1 tube blow out on me since I never run the bias rich. 50-55mA instead of pushing 65.
There are other ways to get richer sound and snappy bass than just biasing up the power tubes (eg. choice of coupling caps and making sure the diff amp tubes are fresh, and adjusted for symmetry with the trimpots)
I still remember fondly the conversations with Leonard and how he got me started with Inifinicap for my old Audio Research amps.
If cost were no object for this project, (hopefully sometime later), I'd be using V-cap teflon 0.22 - 0.33uF for the coupling caps (yes they are *that* good). Anything larger simply wont fit.
My next choice would be Jenseon oils but after having a few oil leak incidents I'd rather stick to the cheaper (but metal casing so use extreme caution) k40y9. Auricaps and Dynamicaps will fit easily, a few notches above the Relcaps, but not enough of an improvement to justify all the work esp on this amp IMO. Too many caps to replace.
All the above have leads which will not fit into the solder holes, you will need to do a 90 degree bend of about 2-3mm of the leads and solder them parallel to the pad. Also be EXTRA gentle and careful not to damage the traces when you are extracting the old caps.
The large can capacitors are always Sprague or Mallory or Cornell Dublier. Easy to swap out if you can find them in stock at Allied or Digikey or Mouser. Buying from Audio Research would be fail-safe of course. |
@Sbrown: Check VHAudio V-cap's web page on sizing coupling caps. The -3dB rolloff point is lower than audible. Plus there is hardly over 2 inches of traces from plate to grid for each coupling stage.
Also even in the 90s when I used to call Leonard at Audio Research to get suggestions on how to upgrade the Classic 120s he's recommend 0.47 or 0.68uF Infinicaps (Dynamicaps werent available yet) to replace the paralleled 1uF Rel 0.01 Wondercaps.
Larger caps were used in the 70s and 80s to add a thicker sonic signature to mask the mid high glare. |
@Hifigeek1: Those huge lytics are Sprague and Mallory and hard to find in-stock @ Digikey or Allied. The cap plugs are impossible to buy in qty < 1000. |
My experience with stock vintage Audio Research gear is that all of them distort/smears quite early with complex passages esp. chorale and loud piano chords.
I also have a problem with constantly trying to adjust the volume when listening to music. To low and I miss resolution, too loud and the glare is mighty annoying.
Not sure if the Relcap/Wondercaps are not aging well which contribute to that or not.
Everytime after I re-cap these amps with any new caps I often find they get more headroom before they start smearing and exhibit glare.
I have tried Infinicap, Dynamicap, Obligato, and Auricap on my previously owned Audio Research amps with varying levels of success. Each has their strengths and weaknesses. All of them, while better than the Rel/Wonder combo, still exhibit glare IMO.
My old Counterpoints (except for the NPs) all use Wondercap. All have benefitted from changing to oil or Teflon caps.
However, I have not tried the state of the art caps like Mundorf SGO or Vcap or even Jensens on my previous Audio Research Power amps ($$$, number of caps involved) like i have on other gear but based on my experience with the k40y9s on other gear and on this D115Mkii I am confident that this amp can even be further improved with the class AA+ caps. (Audio Research uses Teflon nowadays, wonder who makes them?)
Also regarding the size of the caps, look at what other manufacturers use these days for similar circuits. Very few of them use 1uF or larger anymore. Preamp output caps are a different story.
This is only the first step in modding this unit. I am still trying to find a way to install fast recovery diodes and am considering upgrading the high voltage bypass as well. Not now but later. |
@Hifigeek1: After pulling 8 screen grid Rs and tying pins 3-4 on all 8 power tube sockets and 2 resistors in the bias regulator circuitry (was tuned for pentode mode at 40V I need 49 or so) I got myself an all tube CL60 :)
Whether running it as a triod D70 (half tubes) or D115 this is by far the absolute most delicate and musical sounding Audio Research Power amp I have ever owned.
The Classic series never got close to this level of transparency and involvement. |
@Daveyf: I meant even after replacing the Rel+Wonders with the class B caps like Auricaps there is still glare remaining. Not so this time, I am using oil caps, not polyprop.
My 2 main pairs of speakers are Avalon Eidolons (too big for my 14 x 18ish room) and a highly modified pair of Utopia Divas. The Eidolons with their Accuton ceramic tweeter tend to exacerbate the glare and hardness more so than the Utopia.
I use the 4 Ohm taps.
I never have a glare problem with my Phi70s (well after all its an expensive amp with all 300Bs and Infinicap-S, they MUST not have glare) or my Counterpoint NPs (Vcap TFTF).
With the stock caps, and if you are not driving the amp hard, you wont hear a lot of details in the recordings.
Thats my main complaint of the Rel+Wonder combo... low volume and its muddy, high volume and it pounces at you.
Anything oil or Teflon will convey the microdetails better and be a lot more linear when going from soft to loud. I never got a chance to discuss with Leonard about teflon caps but figure why he recommended Infinicap and later on Audio Research went from Rel+Wonder to Dynamicap and now Teflon (in some of their amps), not sure about the TRT Stealth which should be a modern version of Dynamicaps. |
@Smoffat: The reason why I sold my CL120 (and CL150 of course) was the solid state input and gain stages.
I believe in the Counterpoint (Michael Elliot) philosophy of getting the small signal stage as tubey and transparent as possible and use whatever as the output stage as it has the least influence to the sound.
The 10uF caps had to be output caps (from pre to power) or from phono to pre. Upgrading those to any exotic brand like Duelund or Mundorf or VCap would be cost/size prohibitive.
The coupling caps within each device had to be 2uF or under I imagine. But until I get to see the schematic I cant comment.
Anyway, upgrading the small signal coupling caps would have a more noticeable benefit than the output caps.
|
@DaveyF: It's AOK... to each his own.
I am writing this echoing my opinion which concur with that of many others, esp on the more DIY oriented forums.
Perhaps too many people who posted similar opinions have metal dome or ribbon tweeters that need some balance.
BTW I didnt come up with the idea of replacing the Rel+Wonder combo with Inifinicaps. It was a suggestion by none other than Leonard @ Audio Research. |
I have a batch of 0.1, 0.22, and 0.47 FT3s laying around.
Cant use them anywhere other than my point to point wired projects. Too big.
All my Vcaps are deployed in various pcs of existing gear.
Are yours branded ones or generic Russian? Definitely not the Psvane Audio Teflon leaded caps right?
Tried to send you a PM but cant get around this user interface merry go round (pun somewhat intended relative to the Moncrief merry go round :) ) |
@HifiGeek1: Audio Research seems to source caps from The Reliable Cap company (Audio + Rel) and TRT (which some think is sourced from ASC). I will pass on big sized caps for now. Even the 0.47 I have I couldnt find anything to fit them in.
@DaveyF: For coupling caps I would say V-cap Teflon TFTF or Mundorf Silver Gold Oil. Tried/have them on many friends'gear as well as my own. Never tried the Duelunds, no I didnt win the megamillions lotto. Jensen ceramic caps, maybe, too expensive still. |
@Hifigeek1: I was trying to do some final checks on the regulated voltages to the first stage B+ and regulated screen grids and I found an ultrasonic oscillation of about 20mV peak to peak. My mistake for not turning the horizontal dial of my scope all the way to look at ultrasonics. Thought the 20mV was just hash or the probe.
This was traced all the way back to the plate of the 12AT7, and then back to pin 6 of the TL071 (U1) which was sending 100mV of sawtooth.
I'm leaving this for another day now but am curious if you think I should replace the 3 0.22 caps surrounding the input pins of the TL071 opamp.
Much appreciate any guidance. |
@Hifigeek1: Never mind, the relcap 2.0uF bypassing the cathode (regulated B+) output of the regulator 6550C has deteriorated enough that when I paralleled in another 1uF all is good. Time to replace them bypass Relcaps too I guess. |
@DaveyF: great to hear u are in love with your D70 all these years... much like me --> my Counterpoint NP220 Premium Gold. I am surprised that Olsher didnt mention anything about converting it into triode. I was under the impression that he's all for triode glow. |
Ampex53: I dont have a thread on modding the D115ii.
The schematics are floating around the Internet.
The docs Kalvin sent me are the schematics (same) and parts list.
A 1987 build would be a D115ii.
Hope the fuse holder is repaired by now.
Biasing is a nightmare but since bias V is regulated you can put in 1 row of KT88s and bias them then add the other row. You MUST use DVM probes that have smaller exposed tips otherwise you risk shorting out things.
More importantly is the diffamp balances.
The gain stage 75V plate voltage and the 250V driver plate voltages must match (the 250V is hard to match you need to have 6FQ7s that are very symmetrical within 1 envelop)
Mods that I have done:
* All coupling caps to 0.33uF Russian K40y9 PIO
* Triode mode (pull the screen grid resistors and connect plate to screen grid with 100Ohm resistors @ each KT88(need change out a few resistors @ the bias regulator circuit to raise regulated ference bias voltage to close to 50V otherwise you smoke the KT88s)
* a very involved mod which involves REMOVING 1 pair of the 6FQ7 (V4V5) and substituting the V6V7 driver 6FQ7 with 6H30. THis involves tying the 2 6kohm (schematic says 5k) R27R28 together to form a common cathode R of 3kOhm and reducing the Plate R and pot to 13k plate resistance. (this is NOT for the faint of heart, I came up with those values running many rounds of spice simulation)
I am extremely happy with the results. |
No worries hifigeek1 I removed the 2 6FQ7s (V4V5) which are the capacitatively coupled bottom tube of the 4 triode criss cross diffamp driver subcircuit so the net filament draw of this new config is 1 pair of 6H30 vs 2 pairs of 6FQ7 ie 30% lower than before.
The 4 6FQ7s previously shared the same 6.3V AC filament winding.
My amp now runs with 2 less tubes and 4 less coupling caps than stock,
6922 phase-split->cap->6922 gain->6H30->cap->KT88. |
yes and yes.
6922 plate remains 75V as prescribed.
I did not change the 1 1/2 6922s for each channel.
Just REMOVED the first 6FQ7 and reconfigured the 2nd as common cathode.
Note that in the orig circuit the 6922 with plate = 75V is directly coupled to the grid of the 2nd (driver 6FQ7 anyway)
6H30 plates are now 280V (B+ still 455V)
I can send you a spice file. |
