National Semiconductor released a top class audio operational amplifier again:
http://www.national.com/pf/LM/LME49990.html#Overview
http://www.national.com/ds/LM/LME49990.pdf
Who have heard them already?
Thank you for your comments.
http://www.national.com/pf/LM/LME49990.html#Overview
http://www.national.com/ds/LM/LME49990.pdf
Who have heard them already?
Thank you for your comments.
Last edited:
looks interesting spec-wise, but also looks like these parts are not quite "generally available" yet ...
mlloyd1
mlloyd1
Nice specs indeed.
What do we think about the suggested output compensation network as showed on page 9 from its datasheet?
What do we think about the suggested output compensation network as showed on page 9 from its datasheet?
I swapped the LME49720 for the JFET LME49990 in my phono preamp today (basic single chip active feedback design) and it is so much quieter with the new LME49990, Can someone please explain to me why this would be so? I have not played enough vinyl yet to tell if it sounds sonically much different nor have I compared it to my 2 stage passive EQ phono preamp I built with the LME49710HA. Is anyone else experimenting with the LME49990 for phono pre amp's?
My brain has been in neutral most of the day today, It was the LME49880 JFET chip I tried today (not the LME49990). I will try the LME49990 tomorrow.
The cartridge is a Ortofon MC-30 with an Ortofon T-30 transformer, so the preamp is seeing a 47K input. It is the basic circuit as shown on some of the older NS data sheets.
The cartridge is a Ortofon MC-30 with an Ortofon T-30 transformer, so the preamp is seeing a 47K input. It is the basic circuit as shown on some of the older NS data sheets.
LME49720 has very high current noise which makes it a poor choice for the rather high source Z of a phono catridge. A JFET input stage has very low current noise.
Samuel
Just downloaded the LME49880 datasheet - what a hatchet job! The many and various distortion plots show around 0.02% THD at 20kHz - methinks someone didn't factor in the 101 factor of their distortion magnifier... Either that or the headline distortion figures are marketing fluff.
As for the output compensation network - they don't explain why its necessary, apart from improving settling time. Bizarre. What's getting into National these days? this datasheet really isn't up to their usual professional standards, doesn't build confidence for their parts.
As for the output compensation network - they don't explain why its necessary, apart from improving settling time. Bizarre. What's getting into National these days? this datasheet really isn't up to their usual professional standards, doesn't build confidence for their parts.
This is what happens when you lay off the "institutional memory" of a large corporation.
In my real biz we are seeing this with all sorts of huge companies -- they lay off a foreman or plant manager who had been there for 30+ years and then have all sorts of manufacturing issues -- I've been a close-observer of economic and business cycles for almost 40 years and there's something particularly nasty and ugly about this one with how its gutting middle management, where rubber meets the road. Different from 1971, 1982, 1991 etc. The guys at the top (accountants, washed-up Fortune 500 or McKinsey guys) have been parachuted in and have no sense of the organization. All you have to do is look at where the exec resides vis a vis where the company does business -- GM is exempla par gratia.
Who in their right mind would lay off Bob Pease?
Surely this must be getting close to the limits of what's possible re. noise, distortion, PSRR?
Reminds me of the distortion battle of Jap amps in the 70's and 80's though - lower distortion does not always equate to better sound.....but in this case maybe it does.
I'm thinking team it with the LME49600 or a good diamond buffer in the feedback loop and bingo!!
Reminds me of the distortion battle of Jap amps in the 70's and 80's though - lower distortion does not always equate to better sound.....but in this case maybe it does.
I'm thinking team it with the LME49600 or a good diamond buffer in the feedback loop and bingo!!
This is not the first National Semi. 'hatchet job' in recent memory. How about the early datasheet for the audio buffer that had almost the full text and plots of the LM4562 datasheet.
The LME49990 datasheet was written by someone who obviously is not good with the English Language. I can deal with that as not every one of my coworkers at my high-tech company is the best at English (even those for whom is is their mother tongue). But surely some technical writer should proof these marketing tools and fix them up. Aside from the mystery output circuit, there are the fairly specific but vague-on-details bypass recommendations that are not likely good suggestions. Also conspicuously absent are any GBW plots (which might explain the need for the output network). As with all the recent National audio amp datasheets, the distortion plots are meaningless as they plot the residuals of the (Audio Precision?) test set. Surely someone could rewrite the stock LabView program that is churning these out to make some more meaningful plots.
One can only hope that the reality of the device is better than the datasheet and that DIYers in forums like this will chronicle their use of the device: what works and doesn't work, etc., so that we can all benefit.
The LME49990 datasheet was written by someone who obviously is not good with the English Language. I can deal with that as not every one of my coworkers at my high-tech company is the best at English (even those for whom is is their mother tongue). But surely some technical writer should proof these marketing tools and fix them up. Aside from the mystery output circuit, there are the fairly specific but vague-on-details bypass recommendations that are not likely good suggestions. Also conspicuously absent are any GBW plots (which might explain the need for the output network). As with all the recent National audio amp datasheets, the distortion plots are meaningless as they plot the residuals of the (Audio Precision?) test set. Surely someone could rewrite the stock LabView program that is churning these out to make some more meaningful plots.
One can only hope that the reality of the device is better than the datasheet and that DIYers in forums like this will chronicle their use of the device: what works and doesn't work, etc., so that we can all benefit.
Indeed, immediately jumped to my eyes as well.
Not really. Distortion would easily go down even further if we allow higher quiescent current (mainly in the output stage). PSRR is necessarily limited at HF if there is no ground pin connection of the opamp--I really wonder why no IC manufacturer ever decided to make an IC which referrences compensation to ground rather than a supply rail. Noise is more difficult to improve, at least for ICs.
Samuel
I have little idea why you think that this should be so--at 40 dB noise gain these opamps surely perform worse than a SYS-2722.
Samuel
Hi,
voltage noise figures are of dominating interest with low-source impedances, current-noise figures dominate at high source impedances. The OP sees the impedance of the MC-cartridge (transformed up by the audio tranny) parallel to the 47kOhms input impedance. Since the transformed cartridge impedance is considerably smallwer than 47kOhms, the OP is looking into a rather low source impedance. So it´s not as Sam suggested the LME49880´s current noise that makes it noisy in this application, but it´s voltage noise. 7nV/sqrHz is a rather too high value for vynil. It should be 4nV/sqr or less to be inaudible, or say to be lower than the vynil´s own noise figure (which can be no better than -70dB afaI remember).
The LME49990´s datasheet is another of National´s row of bad examples of Datasheets. Useless numbers and figures in multiples and very few peaces of flesh.
jauu
Calvin
voltage noise figures are of dominating interest with low-source impedances, current-noise figures dominate at high source impedances. The OP sees the impedance of the MC-cartridge (transformed up by the audio tranny) parallel to the 47kOhms input impedance. Since the transformed cartridge impedance is considerably smallwer than 47kOhms, the OP is looking into a rather low source impedance. So it´s not as Sam suggested the LME49880´s current noise that makes it noisy in this application, but it´s voltage noise. 7nV/sqrHz is a rather too high value for vynil. It should be 4nV/sqr or less to be inaudible, or say to be lower than the vynil´s own noise figure (which can be no better than -70dB afaI remember).
The LME49990´s datasheet is another of National´s row of bad examples of Datasheets. Useless numbers and figures in multiples and very few peaces of flesh.
jauu
Calvin
Well, the poster suggested that his design is quieter with the LME49880 (note that he corrected the initial part no. two posts later)...
Samuel
Samuel
Are these still unobtanium? No search results from the usual suspects in the UK.
Does anyone have a supplier, please, or are we still in the 14 week lead time?
Cheers
Jon
Does anyone have a supplier, please, or are we still in the 14 week lead time?
Cheers
Jon
How about you give this chip a real test, dump the transformer and let it take on the cartridge directly - as a AD797 can do very well.
Regards, Allen
- Status
- Not open for further replies.