Can I also use the plastic packaged version of the LM3886 with the white thermal goo and the 4U Dissipante? I am not sure how comfortable I feel to mount six LM3886s and avoid shorting them through the metals screw by accident.
Thx
DH
The Modulus-86 is 30 mm tall if you cut the pins tight to the board. I'd plan at least 35 mm in height.
There's a separate thread for the Modulus-86. If you have questions about it specifically, I suggest asking there as it makes it easier for me to get the context and provide a more useful answer. You can find it here: Modulus-86: Composite amplifier achieving <0.0004 % THD+N.
Tom
There's a separate thread for the Modulus-86. If you have questions about it specifically, I suggest asking there as it makes it easier for me to get the context and provide a more useful answer. You can find it here: Modulus-86: Composite amplifier achieving <0.0004 % THD+N.
Tom
Absolutely! I am specifying the plastic package version by default.
Those who are looking to run the amp at ±36 V with a regulated supply to squeeze every availably pW out of it, should change to to the metal back version of the LM3886. They'll have to assume the risks involved with that.
The finished modules that I provide will have the metal back version of the LM3886 so people can push the modules if they'd like to.
Tom
I need some TKP DH2 connectors for the power sensing terminals on the Meanwell SMPS but can't find a supplier in the Uk that stocks them. Nothing on eBay either. Does anyone know of a supplier in the UK? I could get them from Mouser but it would be crazy to pay their UK delivery charge for just these parts.
I think they're a fairly universal 2.0 mm pin pitch type of connector. Nothing fancy. Is there nothing else available with that pin pitch that looks like it'll work?
If you have a Mouser P/N for a suitable connector, just toss me an email. I'll throw some in for you with my next Mouser order. I can't promise when that'll be. Late April or early May, probably.
Tom
If you have a Mouser P/N for a suitable connector, just toss me an email. I'll throw some in for you with my next Mouser order. I can't promise when that'll be. Late April or early May, probably.
Tom
Do you think these would fit?
DF3-2EP-2C | Hirose DF3, 2mm Pitch, 2 Way, 1 Row Male Connector Housing | Hirose
DF3-2EP-2C | Hirose DF3, 2mm Pitch, 2 Way, 1 Row Male Connector Housing | Hirose
Feel free to pm me since my subjective experiences are with the Modulus 286, Parallel 86 and Modulus 86. This is the Modulus 686 thread.
Thanks,
Anand.
I don't think so. The Mean Well RPS-400 data sheet says it's a TKP DH2 series connector. TKP's website says those are compatible with the Hirose DF11-series, which is available at Mouser and Digikey.
There are two types of these connectors: Insulation displacement and crimp contacts.
The insulation displacement connectors are like ribbon cable connectors. You stab AWG 26 wires into the plug and push some plastic piece down. This causes little blades to punch through the insulation and make contact with the conductor. The drawback is that you have to use the right gauge of wire. Using the clamp tool for closing the connector is usually a good idea as well, though - at least with ribbon cable connectors - you can usually get by with a vise.
I recommend using the crimp contacts and the pre-crimped leads. Basically you buy the connector housing and leads separately. Then just push the leads into the connectors and you're good to go. That's a little more expensive than trying to crimp the leads yourself, but unless you have the $400 crimp tool (yes, they're usually that expensive), it's nearly impossible to get it right. Throw money at problem. Make problem go away.
Pre-crimped wire: Mouser P/N: 798-DF112428SF100724
Connector housing: Mouser P/N: 798-DF11-4DS-2C
It does not look like Mouser has the insulation displacement (IDC) style in their inventory, so that sorta solves that end of the question.
Tom
PS: I've included this in the design documentation just now. Those of you with download links to the documentation should be able to see it if you download the doc again.
Last edited:
Hi Tom.
I might have time to start the 686 earlier than I thought and had a couple questions before I start ordering components. Initially I contemplated to convert my two 286 (3U chassis) but with all the infos you provided regarding heat sink requirements etc I decided to go with two 686 mono blocks with two 4U chassis.
Therefore I wondered once more about the trafo power. You recommend the 500VA for example and it has a nominal current of 10A. Now my speakers dip down to 3ohms (at about 100hz) and you stated in one of the first posts that the 686 can output up to 22A rms. Now the noob question. Can the 500VA deliver such a current for very short peaks or should I rather go with more VA (eg 600 or 700VA) as the cost difference is not that significant given the total built cost?
Secondly I wondered what is the max Power the 686 can sustain driving a 2ohm load? From your post 39 I read that it could be in the vicinity of 100W but it might be much more for peaks? Maybe even a THD + N plot over power into 2ohms, if that would make sense and not too much trouble?
Please don’t rush an answer if you’re still tight up in exams as I am currently traveling and will not be able to order components before I get back.
Thx I really appreciate it.
SH
I might have time to start the 686 earlier than I thought and had a couple questions before I start ordering components. Initially I contemplated to convert my two 286 (3U chassis) but with all the infos you provided regarding heat sink requirements etc I decided to go with two 686 mono blocks with two 4U chassis.
Therefore I wondered once more about the trafo power. You recommend the 500VA for example and it has a nominal current of 10A. Now my speakers dip down to 3ohms (at about 100hz) and you stated in one of the first posts that the 686 can output up to 22A rms. Now the noob question. Can the 500VA deliver such a current for very short peaks or should I rather go with more VA (eg 600 or 700VA) as the cost difference is not that significant given the total built cost?
Secondly I wondered what is the max Power the 686 can sustain driving a 2ohm load? From your post 39 I read that it could be in the vicinity of 100W but it might be much more for peaks? Maybe even a THD + N plot over power into 2ohms, if that would make sense and not too much trouble?
Please don’t rush an answer if you’re still tight up in exams as I am currently traveling and will not be able to order components before I get back.
Thx I really appreciate it.
SH
Last edited:
I appreciate your concern regarding my schedule. My last exam is on the 19th and I'm currently working to finish my last assignment. I'm happy to answer specific questions.
Good choice with the 4U chassis, though you can put a stereo MOD686 into that if you'd like. One MOD686 on each heat sink... Just a thought.
For ±36 V nominal rails, math for 8 Ω load says 593 VA per channel for sine wave and 183 VA per channel for music (14 dB CF). The math for 4 Ω load says 1089 VA for sine wave and 322 VA for music (14 dB CF). These numbers are calculated based on the data sheet information for the LM3886 and assume that here is no supply droop as the load is increased. This is what I've been basing my 350-500 VA per channel recommendation on.
Actual measurements say 360 W into 4 Ω, which equates to about 13.5 A peak. That's within the capability of a 2x25 VAC, 500 VA transformer (recall, 10 A RMS = 14 A peak). Even the 350 VA transformer will be fine as it's only on the transients that the high current is drawn. With a 14 dB crest factor, you're looking at around 6 A RMS with peaks reaching 13.5 A if you crank the amp to clipping with ±36 V regulated rails and 4 Ω load.
I have looked at the THD with 2 Ω load. I'd rate the MOD686 as "2 Ω stable" but am not planning to provide any data for 2 Ω operation. I seem to recall getting up to about 300-350 W into 2 Ω. With the better thermal pads and by bolting my newfangled mounting bracket directly to the heat sink, I may be able to improve on that. I also don't remember if that limit was with the metal back or the plastic package LM3886.
Speaking of mounting brackets: The next batch arrived today. I should have a pair of fully assembled modules in stock within the next few days.
Tom
Good choice with the 4U chassis, though you can put a stereo MOD686 into that if you'd like. One MOD686 on each heat sink... Just a thought.
For ±36 V nominal rails, math for 8 Ω load says 593 VA per channel for sine wave and 183 VA per channel for music (14 dB CF). The math for 4 Ω load says 1089 VA for sine wave and 322 VA for music (14 dB CF). These numbers are calculated based on the data sheet information for the LM3886 and assume that here is no supply droop as the load is increased. This is what I've been basing my 350-500 VA per channel recommendation on.
Actual measurements say 360 W into 4 Ω, which equates to about 13.5 A peak. That's within the capability of a 2x25 VAC, 500 VA transformer (recall, 10 A RMS = 14 A peak). Even the 350 VA transformer will be fine as it's only on the transients that the high current is drawn. With a 14 dB crest factor, you're looking at around 6 A RMS with peaks reaching 13.5 A if you crank the amp to clipping with ±36 V regulated rails and 4 Ω load.
I have looked at the THD with 2 Ω load. I'd rate the MOD686 as "2 Ω stable" but am not planning to provide any data for 2 Ω operation. I seem to recall getting up to about 300-350 W into 2 Ω. With the better thermal pads and by bolting my newfangled mounting bracket directly to the heat sink, I may be able to improve on that. I also don't remember if that limit was with the metal back or the plastic package LM3886.
Speaking of mounting brackets: The next batch arrived today. I should have a pair of fully assembled modules in stock within the next few days.
Tom
Wow. As always - wonderful informative answer. Thx a lot, that will help to make my trafo decision.
1) As for the 2ohm performance. Do you think if the speaker dips to 2 or 3ohms one would get a significant distortion let’s say at about 10W rms? Maybe some small bit of data ?
2) On a different note. Which gain version would you recommend I order, the 20dB or the 26dB? My source (Auralic Altair) can put out 4V into XLRs, so I guess the 20dB version, which enjoys even better N?
3) Any chance to make the Power686 available anytime soon so I can order them along with the Mod686s?
Thx
SH
1) As for the 2ohm performance. Do you think if the speaker dips to 2 or 3ohms one would get a significant distortion let’s say at about 10W rms? Maybe some small bit of data ?
2) On a different note. Which gain version would you recommend I order, the 20dB or the 26dB? My source (Auralic Altair) can put out 4V into XLRs, so I guess the 20dB version, which enjoys even better N?
3) Any chance to make the Power686 available anytime soon so I can order them along with the Mod686s?
Thx
SH
Last edited:
1) I don't recall anything out of the ordinary happening in the THD with 2 Ω load until clipping (300-350 W, AFAIR).
2) See spec table. If you scroll down a bit you can see the data for 20 dB gain as well. Lower gain -> lower noise. Physics!
3) Depends on your definition of "soon". I have the schematic mostly done, but need to think it through before committing to a board run. I expect to have it available towards the mid-to-late May. Maybe closer to 'mid' than to 'late', though I'd rather under-promise and over-deliver.
Tom
2) See spec table. If you scroll down a bit you can see the data for 20 dB gain as well. Lower gain -> lower noise. Physics!
3) Depends on your definition of "soon". I have the schematic mostly done, but need to think it through before committing to a board run. I expect to have it available towards the mid-to-late May. Maybe closer to 'mid' than to 'late', though I'd rather under-promise and over-deliver.
Tom
1&2 great. Thx.
3) Ok. Sounds good. Let me hold off until then and I do a combined order of two Power686 and two Mod686. Please reserve me some of those prewound inductor coils to go along with the Mod686.
Until then I look into the chassis and the backplate cutouts. BTW. Does the Power686 come as an unpopulated board so I can leave out the onboard IEC connector and use the same one I used for my Mod286? This way I would not have to adapt the CAD drawing for the cut outs by much.
Thx
SH
3) Ok. Sounds good. Let me hold off until then and I do a combined order of two Power686 and two Mod686. Please reserve me some of those prewound inductor coils to go along with the Mod686.
Until then I look into the chassis and the backplate cutouts. BTW. Does the Power686 come as an unpopulated board so I can leave out the onboard IEC connector and use the same one I used for my Mod286? This way I would not have to adapt the CAD drawing for the cut outs by much.
Thx
SH
Last edited:
I ordered 250 of them earlier today. Do you think that'll be enough for you?
Only snag: They won't be here until June. I do have about 60 left of the previous batch, so I'm not overly concerned I'll run out before the next bunch show up.Yep. The PWR686 will be a bare board, just like the PWR86 is. All leaded components. I plan to make it with both the IEC inlet and a terminal block. Populate one or the other depending on your needs.
The PWR686 will be larger, so don't cut the holes in the bottom panel just yet.
Tom
I bought my heatsinks from BAL in the UK. They offer them in raw aluminium and black anodised finishes. They heat sinking spec of the anodised finish version is slightly better that the raw version, but maybe that's the black having an effect?
Radiative heat transfer (which is secondary as heat sinks typically reject heat primarily by convection to ambient air) is proprtional to the emissivity of the surface and anodizing can lower emissivity if the color is dark like black. On an overall basis for a typical amplifier this would be a quite minor term. A true radiator say on the ISS would benefit enormously from higher emissivity.
I don't think so. The Mean Well RPS-400 data sheet says it's a TKP DH2 series connector. TKP's website says those are compatible with the Hirose DF11-series, which is available at Mouser and Digikey.
There are two types of these connectors: Insulation displacement and crimp contacts.
The insulation displacement connectors are like ribbon cable connectors. You stab AWG 26 wires into the plug and push some plastic piece down. This causes little blades to punch through the insulation and make contact with the conductor. The drawback is that you have to use the right gauge of wire. Using the clamp tool for closing the connector is usually a good idea as well, though - at least with ribbon cable connectors - you can usually get by with a vise.
I recommend using the crimp contacts and the pre-crimped leads. Basically you buy the connector housing and leads separately. Then just push the leads into the connectors and you're good to go. That's a little more expensive than trying to crimp the leads yourself, but unless you have the $400 crimp tool (yes, they're usually that expensive), it's nearly impossible to get it right. Throw money at problem. Make problem go away.
Pre-crimped wire: Mouser P/N: 798-DF112428SF100724
Connector housing: Mouser P/N: 798-DF11-4DS-2C
It does not look like Mouser has the insulation displacement (IDC) style in their inventory, so that sorta solves that end of the question.
Tom
PS: I've included this in the design documentation just now. Those of you with download links to the documentation should be able to see it if you download the doc again.
There's a DF11 crimping tool on eBay. Only £1095.00