LM3875s & low impedance loads - your thoughts?

[BlackDog]

All,

I have seen on here that some think that the LM3875 is not really suited to driving 4 ohm loads....

I'm wondering what other people's experience of this is, and would be interested to hear what other's think.

-- Especially Peter D..... He is selling a product based on this chip (I think), does he recommend his amp to only be used with 8 ohnm speakers, or does it work Ok in his experience?

Up until recently, I thought my speakers Acoustic Energy AE120 were 8 ohm - the Manufacturer's info states they are 8 - but I read a review of them that said that they are 4 ohm! I haven't done any extensive tests, but a multimeter across the terminals looks more like 4 ohms to me.

So, in my experience, using these speakers and my Premium Brian GT kit, my system sounds very good (to me anyway!) Is it because I stayed with a lower voltage toroid (300 VA dual 18v secondaries)???

Let me know what you think...

SteveM

[Vikash]

The lm3875 is four ohm stable, you can take a look at National's spec sheet which shows the voltages vs load. Yes, because you're using 18v secondaries, you can drive a 4 ohm load better than if you had higher voltages. But drawback is that with 18v, if you drive higher loads (6, 8 etc) then you have less wattage potential.

This is just theory from what I've read, but I'm currently hunting for a 22v transformer which I think is the best tradeoff if I want to use between 4 and 8 ohms.

[officeboy]

Well I don't know what happens when you hit the internal current limit, I would suspect clipping. But an 18Vac supply is right where you stay under that 5A limit at 4ohms.

At 22V, you have not only reached the currant limit, but it's to hot for the TF package also.

[carlmart]

Originally posted by BlackDog
I have seen on here that some think that the LM3875 is not really suited to driving 4 ohm loads....

I'm wondering what other people's experience of this is, and would be interested to hear what other's think.

The problem is really when you have a 4 ohm low efficiency speaker, because then you will have to go higher in voltage and get into the clipping area of the chip.

In such cases you would need to parallel two chips to drive it. But using the Overture Design Guide shows that with a +/-30v supply you shouldn't have any trouble for a 4 ohm load with the 3875. Power should be close to 80 watts.

Be sure to use a 250/300VA transformer though.

Up until recently, I thought my speakers Acoustic Energy AE120 were 8 ohm - the Manufacturer's info states they are 8 - but I read a review of them that said that they are 4 ohm! I haven't done any extensive tests, but a multimeter across the terminals looks more like 4 ohms to me.

So, in my experience, using these speakers and my Premium Brian GT kit, my system sounds very good (to me anyway!) Is it because I stayed with a lower voltage toroid (300 VA dual 18v secondaries)???

First of all I should trust the manufacturer spec, particularly if it's a serious one like AE. If they say it's 8 ohm it should be. Maybe there's some dip in the low frequencies, but most speakers do, particularly if they a bass duct. The static resistance is not really that meanigful. The NHT 1259 I have is an 8ohm type and also reads 4 ohm.

What's the efficiency AE says it is? That data might be important in the review.

Think this: the AE was designed for the British market, which usually prefers low wattage amps and small speakers. Such combination has to be efficient, though usually less efficient speakers sound better. 30 watts should move the AE a lot, so with 80 watts you should be covered.


Carlos

[alexw88]

I built a Gainclone with 18V supply and used 2000uF per rail. I am using it with a pair of 4 ohms speaker and I am very happy with it.

Alex

[Vikash]

Originally posted by officeboy
Well I don't know what happens when you hit the internal current limit, I would suspect clipping. But an 18Vac supply is right where you stay under that 5A limit at 4ohms.

At 22V, you have not only reached the currunt limit, but it's to hot for the TF package also.

The graph shows that the chip peaks at 28v input before dropping. Does this drop signify clipping, or the inbuilt protection or what?

So 22v secondaries after rectification is 31.1v, which is too much according to the chart. (BTW, what voltage loss is there through the rectification process, or does the x 1.414 account for this?)

18v comes to 25.45v, and looks like a sensible voltage to get good wattage out of the chip and remain under clipping (according the graph).

[dhenryp]

I don't think there is anything wrong with driving a 4 ohm load with an LM3875 at 30+ volts as long as you keep the input level low enough so that it does not try to drive the amp beyond its current/thermal capabilities. The graph does not mean that the amp will always clip when driving a 4 ohm load with a Voltage > 28 volts. What it is saying is that the chip can not put out more than 55 watts of continuos power into a 4 ohm load. This is a function of the chips ability to dispipate heat long term. There could, in fact, be advantages to to using a higher voltage with 4 ohms. While you still can't put out more than 55 watts continuous it will allow for much higher short term pulses.

[BobK]

Just a few thoughts on my experience using GC's with 4 ohm loads. I have tried the following Avel transformers with a three way active speaker system. The amplifiers were built as monoblocks - so one transformer was powering one channel.

18V - 160VA
22V - 160VA
25V - 250VA

In each case the higher supply voltage gave less usable low end out of the GC's - and produced substantially higher operating temperatures. I even tried 10,00uf caps on each rail of the 22 and 25V supplies and they still clipped before the reference 18V system.

The woofer section presented it's lowest impedence of about 3.2 ohms at 130Hz. If as some have said the 3875 limits at around 5 amps then my max voltage swing is around V=IR or about 16volts.
I assume this is rms current which would put my expected voltage max at about 22.6 volts peak. I don't have a storage scope but on my old HP unit I observed signals clip in the 22 to 25V peak range with the 22V trans and no 10Kuf caps.

I'm going back and try the 10,000uf caps with an 18V/250VA transformer and will let you know what I find.(Couple of weeks before I can get to it)

Just one other thought - the heat dissipation was enough that the 18V ref. unit had no heatsinking other than the case, while the 22/25V units had heatsinks and still got much hotter.

[BlackDog]

Thanks for your replies people, keep it coming !!

Without doing some further tests, I can't really be sure of the impedance of my speakers now! I'm tending to believe the manufacturer's claim that they are 8 ohm.... the sensitivity is claimed to be 89dB @ 1W.

Now, I don't tend to play music REALLY loud these days, but I had to test the amp, didn't I??

It looks as if I'm unlikely to go into clipping with this setup - I found some music that I like, with a strong output, and turned the volume up, and up, and up. I actually had it running at full power for a short while. That kind of level is way beyond "comfortable" listening levels, but still there was no sign of clipping or distortion. Obviously, things weren't as smooth at that level, but hey, I didn't expect them to be...
I did sort of expect the bass to run out of puff, seeing as there's only 1500uF per channel of capacitance to help it along..... still wrong ! (it is possible that the caps on the other (unused) channels of this 6 channel amp were helping out though..)...

Maybe there's some dip in the low frequencies, but most speakers do, particularly if they a bass duct.

Yeah Carlmart, these spkrs have 3 reflex ports each. I bought them back when I was living in England, and used to drive them with a 35 watt amp (AMC 3025a), which also produced enough volume for my tastes. BTW, I am using 2 x 300 VA 18v toroids. When playing in simple two channel stereo, each channel has a 300VA transformer to itself....so no current problem there !

SteveM

[carlmart]

Originally posted by BlackDog
BTW, I am using 2 x 300 VA 18v toroids. When playing in simple two channel stereo, each channel has a 300VA transformer to itself....so no current problem there !

18v + 18v will not provide you quite a high power, even if you use 1000VA transformers, be aware of that. So I am glad you did lisening tests and found not audible clipping, as you should if they clipped. It means efficiency is fine.

On the tests I did, using 3886s which are more efficient, the 18+18 amp clipped before than another using 30+30 in 8ohm speakers that only diped to 6 ohm. The former clipped at higher levels, probably because speakers were not too efficient.


Carlos