Jump to content

and now for something completely different part 3


kevin gilmore
 Share

Recommended Posts

5 hours ago, MLA said:

So around 2A in total (pos and neg rail combined)? If I remember correctly, Dynahi pulls about 2.4A in total (600mA per rail per board), so you'll be close to Dynahi levels with your next step

Btw, can GRLV deliver that much current? I thought it maxed out at 1.5A, but maybe that was per rail, not total?

That is correct - about 1 A per rail so 2 A both rails combined if the bias current is set to 200mA.

I am waiting for parts to finish the GRLV for this amp so will report back later. I am using  a GRLV to supply around 1A current per rail for my DAC but it's only running 15VDC rails and it has not blown up yet... 

Link to comment
Share on other sites

I could be wrong, but I don't think that you combine the power draws of each rail? I think Amb explained it once that the draw is actually rail to rail and not each rail to ground. As I said, I could be misremembering this [emoji4]

 

 

Sent from my iPhone using Tapatalk

 

Link to comment
Share on other sites

25 minutes ago, Pars said:

I could be wrong, but I don't think that you combine the power draws of each rail? I think Amb explained it once that the draw is actually rail to rail and not each rail to ground. As I said, I could be misremembering this emoji4.png

 

I wonder if this depends on the design of the PSU? 

IIRC, Kevin's Golden Reference HV and LV are both two regulators tie together to form a bipolar PSU so measuring each rail separately seems to make sense to me. But I could be wrong about this. Maybe Kevin and others will chime in... 

Edited by mwl168
Link to comment
Share on other sites

19 hours ago, mwl168 said:

I am using  a GRLV to supply around 1A current per rail for my DAC but it's only running 15VDC rails and it has not blown up yet... 

So, presumably 1.5A per GRLV then (i.e. 3.0A for a bipolar one).

That said, Sören’s suggestion above is very interesting! I’ll think about rebuilding my Dynahi Psu that way, just to try it out.

Link to comment
Share on other sites

19 minutes ago, gepardcv said:

I was thinking about how to shrink the PSU for a small-chassis CFA build, and remembered @Kerry's mini-Dynalo. He used two Mean Well IRM-15-24 modules. Would that work for a 24V CFA?

That was actually what I was thinking as well (got boards on order) :D. The 20W IRMs are the same size as the 15W modules by the way and they should be fine for SE. For BAL the 30W ones might be needed but they aren't much bigger.

 

//UFN

Link to comment
Share on other sites

I'm not actually sure I'll have any spares - have a couple of local people interested as well. In any case it hardly seems worth it because these boards are small enough that they can be squeezed in under e.g. Itead's 10x10 cm size limit which means you can get 10 boards for $20 excl. shipping.

There are several other cheap options as well for boards this size, e.g. Seed Studio etc. So if you are happy with "standard" quality boards and willing to wait a little then it's probably cheaper and easier to order yourself :D

Link to comment
Share on other sites

Oh, I already have CFA boards... I thought you meant that you already ordered boards for the IRM PSU. :) If no one has those, I'll probably sit down and lay them out myself within the next few weeks, but I'm not very experienced at PCB design, so that might take a few tries before I get it right.

Link to comment
Share on other sites

Close to finish casing my balanced CFP amp. Waiting for a 4-pin XLR socket.

This one runs on 26VDC rails and 220mA bias current. 

I am posting an unfinished amp because I want to report that I tried OPA551 for the servo OP and it appears to work fine at less than half the cost of OPA445 but only good to 30VDC. something to consider for those that may be in the process of sourcing parts for their builds.

The bias current proves tricky to dial in. It drifts downwards as the amp warms up and seem to continue to go lower even after a hour of operation. The range of drifting is significant too. I am observing 70mA and possibly higher.

 

CFP amp 1.JPG

CFP amp 2.JPG

CFP amp 3.JPG

  • Like 12
Link to comment
Share on other sites

5 hours ago, gepardcv said:

Great work.

Did you find that the bias current drifts at lower rails voltages also?

Thanks!

The bias current drifts at 22VDC rails sitting in open air as well. I do think that it drifts more with higher supply rails and with all the guts in the chassis (re: higher running temperature).

By the way, although this chassis makes for a compact and neat build,  I would suggest using a larger chassis unless you like your amp to run rather warm.  The heatsinks for the amp boards are adequate (45 C after 5 hours) but I think the pass transistors of the GRLV are putting out lots of heat - case temperature is about 55 C bolted to the bottom plate. It probably has to do with my stupidity - I am using a 28vac trafo so the pass transistors have to drop 12VDC and provide over 1A current. I should probably have opted for a trafo with 24 vac secondaries. (EDIT: I should mention that the Antek AS1228 secondaries (28 vac) sag to about 26.5 vac with this heavy load and with that input the GRLV could not sustain regulation for 28VDC rails as I originally planned. I am not certain how high a regulated output a 24 vac trafo can provide.)

This amp rekindles my love for the LCD 2.

Thanks Kevin for yet another great amp that seems curiously overlooked by many. And thanks to congo5 for leading me to it and all the help and advices through my build.   

 

Edited by mwl168
  • Like 5
Link to comment
Share on other sites

Michael,

What are you using for insulators for the pass transistors on the GRLV? I was told to use the aluminum oxide insulators (4180G), but the ones you are using appear to be thinner?

Reason I ask is I was blind drilling/tapping the heatsinks for these to switch the DynaFET/Hi over to a GRLV, and screw length is an issue using the aluminum oxide ones. If I could get away with something thinner, a 3/8" screw would be OK. 1/2" bottoms out, and 7/16" are hard to come by locally.

Link to comment
Share on other sites

2 hours ago, Pars said:

Michael,

What are you using for insulators for the pass transistors on the GRLV? I was told to use the aluminum oxide insulators (4180G), but the ones you are using appear to be thinner?

Reason I ask is I was blind drilling/tapping the heatsinks for these to switch the DynaFET/Hi over to a GRLV, and screw length is an issue using the aluminum oxide ones. If I could get away with something thinner, a 3/8" screw would be OK. 1/2" bottoms out, and 7/16" are hard to come by locally.

Kevin and Birgir already answered but I am using standard silicon insulators for the GRLV pass transistors. 

Different story for the GRHV where aluminum oxide insulators and shoulder washes are a must otherwise arcy sparky!

Link to comment
Share on other sites

Have the rest of the parts incoming for an SE version of this. Already had many of the resistors in PRP/Roederstein, a few Dale. Had to order the oddballs (110, 560, 7K, 68K). Didn't have enough output transistors either. Or 5K pots (was tempted to parallel a 10K with a 10K resistor, but ordered some 5K :eek:

I replaced the sigma22 I had with a GRLV I had setup in my DynaFet/Hi setup, and will use this as the testbed. GRLV is set for +/-30V; might drop that a bit, though I do have OPA445s to use.

Michael: I presume the 5K pot is what sets the bias? Did you preset yours to middle? Didn't want the thing coming up biased at 2A or anything :)

I'll probably have this ready to test (and maybe listen to) this weekend.

Did you notice much difference between your initial SE and final balanced configurations sound-wise?

Thanks, and once again, nice job!

Edited by Pars
Link to comment
Share on other sites

Yes, the 5K pot sets the bias current. Turing it clockwise raises the bias current.

I left mine untouched  from how they were shipped (my experience is they typically will sit around the middle of its adjustment range) when I first powered up. IIRC, they did not start up at anything crazy although I was only running 22VDC rails at the time. Those MJF1503x parts are 36W devices so as long as you have proper heatsink (it's a must) they should be fine. It's probably advisable to turn a pot a few turns counterclockwise before powering on to be safe.  

I do like the balanced setup more than I do the initial SE setup. The balanced amp seems even faster and more control. But even at its SE form, this amp was an jaw-dropper to me.  

By the way, have been tinkering and experimenting and I believe that, everything being equal,  the bias current drift is primarily a function of the case temperature of the MJF1503x devices. As their temperature rises, the current drops. The higher the temperature variance the more the current swing.   

Hope this helps!

 

EDIT: I powered down the CFP amp to measure the 5K pot setting in situ - for 210mA bias on 23VDC rails when warmed up, the pot is set around 1.8K. So I think starting up at 2.5K or higher for the 5K pot is probably a good idea.

Edited by mwl168
Link to comment
Share on other sites

Thanks Michael! This is a 2U Modushop Dissipante case, so the heatsinks should be sufficient. They have been for the DynaFet, and it runs kinda hot. The Fet and Dynahi boards are quick swaps (relatively), and the CFP will be as well (same angle mounts and all connectors are the same).

Mini rant: I had used a single pin JST connector for the outputs on the Dyna boards, since they were only setup for the hot out on the board. Digikey no longer stocks the single pin. I could order some more directly from JST, but shipping is kind of high. Oh well, either pay up, or use something else.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

×
×
  • Create New...