[bcowen]

I recently took some repeat measurements of the red Airmid, inputs were the Russian 6S2S (6J5GT) here are the results if anyone is interested in this type of stuff.

Russian 6AS7G
Output Z - 50ohm
1mW into 300ohm - 0.11% THD
1mW into 120ohm - 0.22% THD
1mW into 80ohm - 0.31% THD
1mW into 32ohm - 0.67% THD

Tung-Sol 5998
Output Z - 30ohm
1mW into 300ohm - 0.032% THD
1mW into 120ohm - 0.094% THD
1mW into 80ohm - 0.15% THD
1mW into 32ohm - 0.38% THD

Frequency response at the two extremes.

300ohm



32ohm



1mW into 300ohm square waves.

1kHz



10kHz



100Hz



1kHz sine wave into 300ohm dummy load just before the onset of clipping. So just about 1W out.

Pardon my oscilloscope ignorance, but what do the sloping waves at 100Hz indicate (versus the flat waves at 1kHz and 10kHz)?

 

[L0rdGwyn]

Pardon my oscilloscope ignorance, but what do the sloping waves at 100Hz indicate (versus the flat waves at 1kHz and 10kHz)?

Sure! The sloping of the waves gives info on high and low FR. The curved (as opposed to sharp) leading edge of the 10kHz square wave indicates HF rolloff. The downward slope of the 100Hz square wave indicates LF rolloff. It gives some quick qualitative info on FR, but also will show ringing in the circuit, like a bell being rung by the test signal, and often will tease out any potential oscillations that could lead to amplifier instability.

 

[L0rdGwyn]

Here is a good article on square wave testing.

https://sound-au.com/articles/squarewave.htm

 

[bcowen]

Here is a good article on square wave testing.

https://sound-au.com/articles/squarewave.htm

Thanks!

 

[L0rdGwyn]

Bronze Airmid is away to a friend of Zach's, not sure if he has a Head-Fi account, I guess we'll find out!

I finished up the work on that Zenith radio FINALLY. But it needs an alignment and I'm not buying an RF signal generator, sent it to a radio dude in Dayton today to do the job for me.

Been loving the Forte IV, they sound great, excellent clarity, body, and totally fill the room. The fancy grills I concocted are in production, here is how they will look.



Now that I'm caught up on Airmids, taking a break from doing any hardcore amp building for a little bit, got some life stuff to do and need to catch up on some small projects that have been bothering me (e.g., replacing the terrible sockets in my curve tracer).

Always planning future designs too. Since I have some new headphones on the way, and I am going to be Airmidless soon, I am going to need a headphone amp. My 6A5G amp is always available, but it is truly a speaker amplifier. The 841 hybrid headphone amp is going to be broken down soon (another project), so that might become a spud that I have drafted, would be a relatively easy project.

There is one other idea I am playing with though that is a little more ridiculous, an OTL using the Airmid topology but using a pair of 6336 or 6528. Might just be a flavor of the week type of thing, but I am seriously considering it. I haven't seen a headphone OTL that uses these tubes at their high current operating points - the 6528 has a gm of 37 mA/V per section, so running them in parallel would bring the theoretical output impedance to something like 15ohm. This would have to be a whole new design though in terms of layout / chassis as doubling the current through the regulators, tubes, and active loads poses some heat management challenges. Would consider active cooling using silent Noctua fans with intake below the tubes through the bottom panel. Would probably do something crazy like 1200uF of film output capacitance too...

We'll see, might have another wild idea tomorrow. I am on a horrible string of work shifts, eight of of nine days, night shift, thirteen hour shifts, brutal, so maybe I'm losing it lol but I think this OTL would be pretty interesting. Will have some nice ZMF headphones at the end of the week though when I'm done with work, can't wait.

 

[Monsterzero]

Airmid topology but using a pair of 6336 or 6528.

Im not familiar with a 6528, but wouldnt an OTL that uses 6336s be rather limited when it comes to tube choices? I have two different 6336 pairs, Cetron and another Chatham (iirc) and to my ears , neither are anything special sounding.

 

[L0rdGwyn]

Im not familiar with a 6528, but wouldnt an OTL that uses 6336s be rather limited when it comes to tube choices? I have two different 6336 pairs, Cetron and another Chatham (iirc) and to my ears , neither are anything special sounding.

It would be limited to the 6336 and 6528. In a circuit that is optimized for them and pushing 300mA each, they would sound very good

 

[gibosi]

If I understand you correctly, while the Glenn can use the 6336 and 6528, the amp is designed around the 6AS7, and isn't optimized to run these high current tubes. I believe the Glenn pushes only 100mA through them versus 300mA per your design. If you do this, I bet it will be a beast! And I, for one, hope you go for it!!

 

[bcowen]

It would be limited to the 6336 and 6528. In a circuit that is optimized for them and pushing 300mA each, they would sound very good

I always get ticked off when I see the Ebay ads throwing the 6528 and 6336 in as equivalents to 6080's and 6AS7G's. Some poor newbie is going to take that at face value and end up frying their amp with twice the heater current.

I'd love to hear an amp optimized for them though. They're half (or less) the going price of the top tier 6080's and 6AS7's.

 

[Monsterzero]

I always get ticked off when I see the Ebay ads throwing the 6528 and 6336 in as equivalents to 6080's and 6AS7G's. Some poor newbie is going to take that at face value and end up frying their amp with twice the heater current.

I'd love to hear an amp optimized for them though. They're half (or less) the going price of the top tier 6080's and 6AS7's.

Glenn suggested the 6336 for driving low Z headphones, so I called Andy over at Vintagetubeservices and he mustve asked me 4-5 times if I was sure my amp could use them. He also refused any type of warranty on the 6336s.

Ive since learned that headphones like the K701 sound pretty spectacular on the GOTL, regardless of which power tube is in there.

 

[Monsterzero]

If I understand you correctly, while the Glenn can use the 6336 and 6528, the amp is designed around the 6AS7, and isn't optimized to run these high current tubes. I believe the Glenn pushes only 100mA through them versus 300mA per your design. If you do this, I bet it will be a beast! And I, for one, hope you go for it!!

It would be limited to the 6336 and 6528. In a circuit that is optimized for them and pushing 300mA each, they would sound very good

What would the sonic benefits be of pushing the 6336 to full capacity, and would those benefits outweigh missing out on some of the more legendary 6as7g and equivalents?

 

[gibosi]

What would the sonic benefits be of pushing the 6336 to full capacity, and would those benefits outweigh missing out on some of the more legendary 6as7g and equivalents?

You may just have to get in line to get one of these amps to find out.

And there might be more tube choices than just Chatham / Tung-Sol / Cetron. I'm not sure, but I think that Raytheon also manufactured these tubes. I have a pair labeled Raytheon that don't sound like my Chathams. To my ears, the Raytheons are not as good. And there is also a Thomson - CSF 6336A, made in France. Like the Thomson 6080, vocals are a bit more forward than the Chatham, but I like it with darker drivers and/or rectifiers.

Below, Chatham to the left and Thomson to the right.

 

[L0rdGwyn]

I always get ticked off when I see the Ebay ads throwing the 6528 and 6336 in as equivalents to 6080's and 6AS7G's. Some poor newbie is going to take that at face value and end up frying their amp with twice the heater current.

I'd love to hear an amp optimized for them though. They're half (or less) the going price of the top tier 6080's and 6AS7's.

Yeah, the circuit needs are very different! 2x the heater and bias current. Not easy to do in a small package, likely a reason they aren't being used in a lot of headphone designs, will get scorching hot too. The price differential is also appealing, given a fully biased 6336 has a similar gm to a 5998 at 1/4 the price.

What would the sonic benefits be of pushing the 6336 to full capacity, and would those benefits outweigh missing out on some of the more legendary 6as7g and equivalents?

The theoretical benefit is a higher transconductance operating point, which in a cathode follower circuit means lower output impedance and distortion. Hard to say exactly how they compare as I have not heard them side-by-side, but I do feel there is a correlation between increasing transconductance and increasing soundstage / clarity / detail retrieval in cathode followers. Transconductance increases as one goes from 6AS7G < 5998 < 7802. A fully biased 6336 has similar transconductance to a 5998, whereas a fully biased 6528 has a transconductance a little over 2x that of a 5998.

You may just have to get in line to get one of these amps to find out.

I'm sorry but not an amplifier I am offering for purchase, just a theoretical DIY project that might come to fruition. I hope what I post here entertains and maybe inspires other people to get into DIY. I might choose to do one-off builds for Head-Fiers, but not open for business, so to speak, I just don't have the capacity for it.

 

[L0rdGwyn]

I've made moves to build the the 6336 / 6528 OTL, it's happening and it's going to be a beast.

Given the power tubes require twice the heater and bias current, two of the mains transformers used in the Airmid are needed, one for each channel. The B+ regulators will be dissipating around 12W each, so they will be mounted on a 200mm x 80mm x 40mm heat sink on the rear panel.

The real challenge comes with the increased bias current and power dissipated in the output tube loads. With the sections paralleled, each 6336 / 6528 will be biased at 300mA, as such the load for each tube will need to dissipate around 34W as they will be dropping roughly half the B+ voltage...LOL.

How to get that done? Big ass heat sinks. Going to put big finned heat sinks on the sides of the chassis, 300mm x 80mm x 40mm, courtesy of the diyAudio store.



At 300mA bias, each power tube will be dissipating 24-30W. They will get hot as hell, so I am going to actively cool them. I already have a 5V regulated supply powering the time delay circuitry, I am going to beef up the LDO regulator in that position for higher current and add 3-pin fan headers to the PCB. Will then run a pair of 80mm Noctua NF-A8 5V fans with intake directly below the power tubes.



Output caps are going to be 600uF 450V ClarityCap TC2. They are huge caps, 77mm diameter by 145mm length. Don't have a pic, but will look something like the middle cap below.



I have a general idea of the layout, but need to get the parts in, take some measurements before I can finalize. No prototype, just going to build it.

 

[bcowen]

I've made moves to build the the 6336 / 6528 OTL, it's happening and it's going to be a beast.

Given the power tubes require twice the heater and bias current, two of the mains transformers used in the Airmid are needed, one for each channel. The B+ regulators will be dissipating around 12W each, so they will be mounted on a 200mm x 80mm x 40mm heat sink on the rear panel.

The real challenge comes with the increased bias current and power dissipated in the output tube loads. With the sections paralleled, each 6336 / 6528 will be biased at 300mA, as such the load for each tube will need to dissipate around 34W as they will be dropping roughly half the B+ voltage...LOL.

How to get that done? Big ass heat sinks. Going to put big finned heat sinks on the sides of the chassis, 300mm x 80mm x 40mm, courtesy of the diyAudio store.



At 300mA bias, each power tube will be dissipating 24-30W. They will get hot as hell, so I am going to actively cool them. I already have a 5V regulated supply powering the time delay circuitry, I am going to beef up the LDO regulator in that position for higher current and add 3-pin fan headers to the PCB. Will then run a pair of 80mm Noctua NF-A8 5V fans with intake directly below the power tubes.



Output caps are going to be 600uF 450V ClarityCap TC2. They are huge caps, 77mm diameter by 145mm length. Don't have a pic, but will look something like the middle cap below.



I have a general idea of the layout, but need to get the parts in, take some measurements before I can finalize. No prototype, just going to build it.

Holy cow. Are you also going to need a dedicated 30 amp circuit for it?