I built LTSpice models for the NTD1 and combined NTD1 and BAL-BAL to simulate the different resistor and voltage options for a DC-coupled NTD1. Thought I’d share the models and results in case anyone else is interested in experimenting:
NTD1 23/49 (0.25VRMS) + BAL-BAL (4x gain) @ 12V & -6dBFS
NTD1 30/65 (0.33VRMS) + BAL-BAL (3x gain) @ 12V & -6dBFS
NTD1 45/98 (0.50VRMS) + BAL-BAL (2x gain) @ 18V & -6dBFS
My plan is to go with the 23/49 NTD1 and 4x BAL-BAL, which seems to simulate the best, and power both boards with a pair of Parallel LT3042 supplies.
There was lots of talk of transconductance in the original NTD1 thread, with the general consensus being that higher transconductance and higher current led to better performance.
The NTD1 BOM calls for the Fairchild FDB52N20, which Owen says is the best he’s found at 100mA. Since the 23/49 NTD1 will have about 280mA running through each leg, are there MOSFET characteristics I can look at to see if there’s a part with higher transconductance at this operating point?
Few data sheets have a transconductance graph (the ones that do are next to impossible to read at low current) and typical gfs is almost always measured at a much higher current. Is this at all discernible from a datasheet, or is building a rig to measure transconductance required?
And this may be a dumb question, but when Owen says that the BAL-BAL current draw is 110mA-530mA from each “supply rail”, in the case of a dual-mono supply does that mean 530mA max from each of the + and - rails (1.06A per channel, 2.12A per board), or 530mA max per channel and 1.06A per board?
NTD1 23/49 (0.25VRMS) + BAL-BAL (4x gain) @ 12V & -6dBFS
NTD1 30/65 (0.33VRMS) + BAL-BAL (3x gain) @ 12V & -6dBFS
NTD1 45/98 (0.50VRMS) + BAL-BAL (2x gain) @ 18V & -6dBFS
My plan is to go with the 23/49 NTD1 and 4x BAL-BAL, which seems to simulate the best, and power both boards with a pair of Parallel LT3042 supplies.
There was lots of talk of transconductance in the original NTD1 thread, with the general consensus being that higher transconductance and higher current led to better performance.
The NTD1 BOM calls for the Fairchild FDB52N20, which Owen says is the best he’s found at 100mA. Since the 23/49 NTD1 will have about 280mA running through each leg, are there MOSFET characteristics I can look at to see if there’s a part with higher transconductance at this operating point?
Few data sheets have a transconductance graph (the ones that do are next to impossible to read at low current) and typical gfs is almost always measured at a much higher current. Is this at all discernible from a datasheet, or is building a rig to measure transconductance required?
And this may be a dumb question, but when Owen says that the BAL-BAL current draw is 110mA-530mA from each “supply rail”, in the case of a dual-mono supply does that mean 530mA max from each of the + and - rails (1.06A per channel, 2.12A per board), or 530mA max per channel and 1.06A per board?
Not sure if you noticed but I did setup the "Wire-amp"(LME49830/LatFET) BOM in Mouser, so others can order from it, if they choose too.
Mouser does not sell the LatFET's, so they will have to be ordered from other distributors.
Missing the TO-264 insulators, so will lok to see what Mouser sells and add them on.
"The Wire AMP" Class A/AB Power Amplifier based on the LME49830 with Lateral Mosfets
Mouser does not sell the LatFET's, so they will have to be ordered from other distributors.
Missing the TO-264 insulators, so will lok to see what Mouser sells and add them on.
"The Wire AMP" Class A/AB Power Amplifier based on the LME49830 with Lateral Mosfets
Got mine in nice casing:
An externally hosted image should be here but it was not working when we last tested it.
Got mine the other day (Cheers Owen! yeah I know you're a busy guy).
All installed in my dual-mono dam1021 build. It's awesome, HD800's are now full of life and so punchy that you can feel them.
When I was using the onboard op-amp buffers of the dam1021's it felt like there was something missing, indeed there was. (lots!)
All installed in my dual-mono dam1021 build. It's awesome, HD800's are now full of life and so punchy that you can feel them.
When I was using the onboard op-amp buffers of the dam1021's it felt like there was something missing, indeed there was. (lots!)
Nice setup spikestabber, got mine years ago, and I've quite a similar configuration (still on a bench table):
Amanero => single dam1021 modded rev1 0.02% raw output => Wire SE-BAL (using mono dual Wire PSU v2) => XLR 4P => Balanced Fostex TH900
My first impression are excellent. As your HD800 is quite the king for headphone soundstage, what do you think about its depth ? did you try playing with the dam filters ?
Amanero => single dam1021 modded rev1 0.02% raw output => Wire SE-BAL (using mono dual Wire PSU v2) => XLR 4P => Balanced Fostex TH900
My first impression are excellent. As your HD800 is quite the king for headphone soundstage, what do you think about its depth ? did you try playing with the dam filters ?
I've settled on the C128dp Linear Phase from Paul for the last several months. Haven't quite found a filter that matches it in preference, its excellent with all kinds of different material and I think its rather neutral in that it doesn't blare everything right in your face. I'm extremely impressed with how it is currently. It's an awesome amp paired with dam1021 raw outputs, bar none.
Bimmer,
You be dah man! Thanks for working with Owen to get these moving again.
See my email for what I've got outstanding.
TIA!
Greg in Mississippi
So... what does availability look like?
If Owen is willing to open-source the design, I would happily run a batch of PCBs and post 'em to people. It'd be good to see design sources on github or similar...
Alternatively (too slow to edit my post!), I'd be happy to draft up some open-source PCBs using the appnote (Figure 28 of the LME49600 datasheet), or following "THE WIRE - FINAL SCH.pdf" if Owen is happy for that happen with his schematic, though it looks like it's a textbook instrument-amp with the LME49600 grafted on the end. The schematic in the datasheet has a DC servo.
Any opinions?
There had been several attempts to implement datasheet schematic of the LME49600 headphone amp with PCB designs. Just search the head-amp forum.
In his early posts Owen discussed the reasons behind omitting the DC servo, DC coupling his amp and other decisions he has made when designed his head-amp. For more details check his SE-SE design thread given in the first post. In principle you do not have to follow his design criteria but then it will be different headphone amp
Regards,
Oleg
In his early posts Owen discussed the reasons behind omitting the DC servo, DC coupling his amp and other decisions he has made when designed his head-amp. For more details check his SE-SE design thread given in the first post. In principle you do not have to follow his design criteria but then it will be different headphone amp
Regards,
Oleg
There had been several attempts to implement datasheet schematic of the LME49600 headphone amp with PCB designs. Just search the head-amp forum.
In his early posts Owen discussed the reasons behind omitting the DC servo, DC coupling his amp and other decisions he has made when designed his head-amp. For more details check his SE-SE design thread given in the first post. In principle you do not have to follow his design criteria but then it will be different headphone amp
Regards,
Oleg
Do you have any links to discussions of problems with the datasheet designs? The relevant threads are up to a couple hundred pages now
I don't have the links to posts discussing other designs and here we discuss "The Wire" family
Regarding Owen's desisions/explanations just look here and here for the start. Just go through the entire thread, 256 pages isn't much if you jump over unrelated posts. Most relevant info is in the early posts anyways, so you'll get the idea pretty quick. Just follow the posts by Owen and then look back if the question answered is not clear.
Regarding Owen's desisions/explanations just look here and here for the start. Just go through the entire thread, 256 pages isn't much if you jump over unrelated posts. Most relevant info is in the early posts anyways, so you'll get the idea pretty quick. Just follow the posts by Owen and then look back if the question answered is not clear.