The Parallel-86 boards arrived today. Of the total order, I have 13 boards left. You can buy yours here: Parallel-86 (Last Chance to Buy)
I'll be getting some LM4780s to go with the boards. You can place a backorder of this combination on my website. The backorders will be filled at the end of August.
Tom
I'll be getting some LM4780s to go with the boards. You can place a backorder of this combination on my website. The backorders will be filled at the end of August.
Tom
In case anyone is still holding out for the Parallel-86 boards: I have five of them left from the group buy. I'll have the LM4780 to go with them in about a week. You can get yours here: Parallel-86 (Last Chance to Buy)
Tom
Tom
tomchr
1. Do you have any new designs with more power planned, and if yes when can they be expected?
2. Will you ever do an amp using the tpa3251d2?
1. Do you have any new designs with more power planned, and if yes when can they be expected?
2. Will you ever do an amp using the tpa3251d2?
I'm working on the Modulus-286. It is an all-SMD build on a 4-layer board of two channels of Modulus-86. I have a prototype running and it performs very well. It has lower noise than the Modulus-86 and has a default gain of 26 dB (20x) and an input sensitivity of 900 mV RMS. The channel separation measures about 90 dB, which is quite good, but I may be able to improve that further.
There will be resistor options to allow the two channels to be coupled in parallel for higher output power.
In stereo mode (±30 V supply), this amp provides 40 W into 8 Ω and 60-65 W into 4 Ω.
In parallel mode (±35 V supply), the amp provides approx. 60 W into 8 Ω and about 115 W into 4 Ω.
I'll start out the boards at $150/each for the final version. I expect to have the final version available in late September. The BOM cost is $90ish.
For those looking to become early adopters and would like to save $50/board in exchange for having to live with a few limitations have a new option: I have 11 of the prototype boards (Rev. 1.0) and will need 3-4 of them for my own use. I'm willing to part with the rest for $100/each. They will work well for stereo builds and operation up to ±35 V (absolute maximum) but will not support parallel operation. In addition, you need to stack two ceramic caps in four places. The final board will have these kinks ironed out, allow parallel operation, and up to ±42 V supply voltage (absolute maximum, ±35 V is the recommended design target).
Early adopters interested in these boards, please toss me an email through the Contact Us page on my website or PM me.
Attached image shows the Modulus-286 Rev. 1.0 prototype in stereo configuration.
In addition to the Modulus-286, I am looking at a discrete composite design with higher output power. It would be pretty neat to hit a headline spec of 200 W (8 Ω) but the thermals involved with such a design are not trivial. More math is needed. That design is probably a project for next spring. It will NOT be a chipamp design and the BOM cost will be significant.
If you are looking for an amp project that will come together easily and require no adjustments, the Modulus-86, Modulus-286, and Parallel-86 (until I run out) are your best options.
I will keep the Modulus-86 going as long as I can as it is my only all thru-hole power amp design. Some builders are still reluctant to take the plunge to SMD. Expect my future designs to include at least some SMD components where required to obtain good performance.
I have no plans to venture into Class D. If you're looking for a good Class D amp, I suggest looking at the Liteamp design that's floating around here or buy a pair of Hypex Ncore modules.
Tom
Attachments
Last edited:
I can't wait for the measurements on this one! Even though I don't need the power, I'm curious nonetheless!
Keep it up Tom!
Anand.
Keep it up Tom!
Anand.
Tom,
Major kudos for an exceptionally well engineered product. I'm an ME, not an EE, so I could only scratch the surface of the technical details that you've undertaken, but I understand full well the painstaking process of your modeling, prototyping and analysis during the design of the Mod-86.
I will need at least a 4-channel setup (possibly 6) to use in an active dipole system similar to an LX521. If I pull the trigger on Mod-86, I will get the fully assembled boards and assemble into an amp chassis that I will source.
I do have a few questions, mostly related to overall amp chassis size requirements and layout:
1. What are the physical X*Y dimensions of the mono Mod-86 and the projected stereo Mod-286 PCB's?
2. Do you have any cost projections for the fully assembled Mod-286? With the higher SMD part count, the assembled cost could potentially be proportionally less for the stereo version depending on run qty (?)
Once I know the sizes of the mono and stereo versions (and projected cost), I'll lock in a specific amp chassis layout and get the Mod-xx6 boards on order.
Thanks again for what looks like a stellar addition to the diy world 😉
-Brad
Major kudos for an exceptionally well engineered product. I'm an ME, not an EE, so I could only scratch the surface of the technical details that you've undertaken, but I understand full well the painstaking process of your modeling, prototyping and analysis during the design of the Mod-86.
I will need at least a 4-channel setup (possibly 6) to use in an active dipole system similar to an LX521. If I pull the trigger on Mod-86, I will get the fully assembled boards and assemble into an amp chassis that I will source.
I do have a few questions, mostly related to overall amp chassis size requirements and layout:
1. What are the physical X*Y dimensions of the mono Mod-86 and the projected stereo Mod-286 PCB's?
2. Do you have any cost projections for the fully assembled Mod-286? With the higher SMD part count, the assembled cost could potentially be proportionally less for the stereo version depending on run qty (?)
Once I know the sizes of the mono and stereo versions (and projected cost), I'll lock in a specific amp chassis layout and get the Mod-xx6 boards on order.
Thanks again for what looks like a stellar addition to the diy world 😉
-Brad
Brad,
Thank you for your kind words. I appreciate it.
The MOD86 board measures 3.550 x 2.725 inches. You need to allow some space for the wiring to get into the terminal blocks, so I'd keep at least 1" of clearance around the board.
The MOD286 board measures 4.450 x 3.150 inches. That is the size of the prototype board and I fully expect the final board to be the same size. All connections are made along one edge of the board. You'll need probably 1.5 inches of clearance there.
Right now I perform the assembly and test by hand, which drives up cost. I haven't figured out the pricing on assembled MOD286 yet, but I figure around $400-450. The board takes a couple of hours to build and test. Once I develop a relationship with an assembly house, I expect that cost to go down.
Tom
Thank you for your kind words. I appreciate it.
The MOD86 board measures 3.550 x 2.725 inches. You need to allow some space for the wiring to get into the terminal blocks, so I'd keep at least 1" of clearance around the board.
The MOD286 board measures 4.450 x 3.150 inches. That is the size of the prototype board and I fully expect the final board to be the same size. All connections are made along one edge of the board. You'll need probably 1.5 inches of clearance there.
Right now I perform the assembly and test by hand, which drives up cost. I haven't figured out the pricing on assembled MOD286 yet, but I figure around $400-450. The board takes a couple of hours to build and test. Once I develop a relationship with an assembly house, I expect that cost to go down.
Tom
Tom,
Excellent, the size info will allow me to do some chassis layout plans. Once you've got the MOD286 dialed in and have final measurements, I'll be ready to pull the trigger on either (four) 86's or a couple 286's. Hopefully the overall volume demand will be high enough to pull in an assy house 😉
Cheers!
Brad
Excellent, the size info will allow me to do some chassis layout plans. Once you've got the MOD286 dialed in and have final measurements, I'll be ready to pull the trigger on either (four) 86's or a couple 286's. Hopefully the overall volume demand will be high enough to pull in an assy house 😉
Cheers!
Brad
Tom,
Any plans for an all SMD version of the Modulus 86 along with mirror imaged boards that measure at least as well as your Revision 2.10? Just curious 😉
Best,
Anand.
Any plans for an all SMD version of the Modulus 86 along with mirror imaged boards that measure at least as well as your Revision 2.10? Just curious 😉
Best,
Anand.
Definitely no plans for mirrored boards as that wrecks performance, in particular of the supply routing. If I make any dramatic changes to the board, I'd probably move the connectors to one edge as I did with the MOD286. That'll eliminate the perceived need for mirrored boards.
I have no planned updates to the Modulus-86 circuit. Right now, Rev. 2.1 is as good as it gets. I am fully committed to supporting the Modulus-86 and delivering boards for a long time to come. If there ever is an up-rev, it will be with automated assembly in mind, hence, all SMD. That said, I have more than enough projects lined up to carry me through the next year, if not more.
My experience with the Modulus-286 will guide the manufacturability aspects. Time will tell.
Tom
I have no planned updates to the Modulus-86 circuit. Right now, Rev. 2.1 is as good as it gets. I am fully committed to supporting the Modulus-86 and delivering boards for a long time to come. If there ever is an up-rev, it will be with automated assembly in mind, hence, all SMD. That said, I have more than enough projects lined up to carry me through the next year, if not more.
My experience with the Modulus-286 will guide the manufacturability aspects. Time will tell.
Tom
Most of the parts are 0805. There are some 1206 and a couple 2510. There are SOIC, SOT-23, SOT-223, and some rather small diodes in SOD-323. Those diodes are the smallest parts.
Tom
Tom
Hello Tom,
How much lower will the noise floor be?
I ask the question because the residual noise of the MOD86 is what has been holding me back from trying it on my compression driver (nc400 has a 25uV unweighted spec, which is already audible on my compression drivers with input shorted).
Regarding gain, will it be easy to adjust to a lower value? How much will it reduce residual noise (internal noise, input shorted)?
Thank you
25uVac is pretty quiet for the noise output of a power amplifier.
that is -101dBW, ref 1W into 8ohms. (= -121dB ref 100W into 8ohms)
Going to a Buffer with no gain from your Source may be a way to get significantly lower noise.
that is -101dBW, ref 1W into 8ohms. (= -121dB ref 100W into 8ohms)
Going to a Buffer with no gain from your Source may be a way to get significantly lower noise.
Don't be afraid of noise in the MOD-86. I just connected mine to my 100dB/W/M Unity horns and it is dead quiet, I mean head inside the horn - no sound - dead quiet. My system has never been this quiet before, despite running very high quality amplifiers and an optimised gain structure.
Make sure your gain structure is good, this can increase your noise floor by >20dB if you get it wrong. I drive mine from a DCX2496 and built my MOD-86 with 14dB (X5) gain - details given in the design docn.
You won't regret this build.
Cheers
William Cowan
He has stated he already hears noise from a specified 25uVac with his set up.
The mod 86 is about 4dB to 5.5dB noisier.
He recognises that gain affects noise output.
+20dB and +26dB gain values of chipamps may be too much for a high sensitivity horn.
30W to 50W of maximum output power may be too much for a high sensitivity horn.
Last edited:
pos,
have a look at the lme49600 datasheet.
there are a few headphone amplifiers that use that topology.
The gain could be set to +0db or +3dB or even +6dB to drive a horn.
The max output of the 49600 is 250mApk and that equates to 250mW into 8ohms.
The datasheet shows the opamp+buffer. What about replacing the 49600 with a diamond driver + ClassA biased EF output. There are probably Threads with this topology.
Potentially you are in the 0.3ppm distortion and noise is ultra low using a modern low noise low distortion opamp like the lme49990.
have a look at the lme49600 datasheet.
there are a few headphone amplifiers that use that topology.
The gain could be set to +0db or +3dB or even +6dB to drive a horn.
The max output of the 49600 is 250mApk and that equates to 250mW into 8ohms.
The datasheet shows the opamp+buffer. What about replacing the 49600 with a diamond driver + ClassA biased EF output. There are probably Threads with this topology.
Potentially you are in the 0.3ppm distortion and noise is ultra low using a modern low noise low distortion opamp like the lme49990.
I want to use a 4 ohm compression driver that is very sensitive (JBL 2435HPL). The mass break point frequency is very high so it is very sensitive up to at less 6kHz, even on a constant directive waveguide.
Thank you for the suggestion.
I have 4 LPHUP power amps here that I could use for that build, but I wanted to get a little more power here, and also be able to drive high impedance loads with enough power (16 ohms 2450SL-Be and 32 ohms D2430K, although this one could be configured as 8 ohms...).
I would be curious to know the residual noise of the 14dB gain version of the MOD86.
Thank you for the suggestion.
I have 4 LPHUP power amps here that I could use for that build, but I wanted to get a little more power here, and also be able to drive high impedance loads with enough power (16 ohms 2450SL-Be and 32 ohms D2430K, although this one could be configured as 8 ohms...).