While looking at making a small headphone amplifier using ok amps and a buffer (5532 and LME49600) I found that I have several buffer opamps (probably discontinued) which are LH0033CG.
The specs don't look too bad.
Has anyone experimented with these buffers for audio? What were your impressions?
Thanks
The specs don't look too bad.
Has anyone experimented with these buffers for audio? What were your impressions?
Thanks
This will mostly depends on your gesign goals and type of headphones.
High-sensitivity ones, including in-ear types, will be driven very well, low-sensitivity ones, especially high-resistance and ortho/isodynamic, could have lack of drive voltage.
What will be an amp core?
Load should be low Z. Probably 32 Ohm or as low as the circuit can handle
This is still an idea that hasn't made it's way to pen and paper yet so I suspect the design will face several limitations on load. As I haven't done any maths yet (and do not expect other to do for me) I am sure load will dictate how useful these ICs may be for a head amp.
It may even be the case that I require a stage following the buffer.
I just thought I would inquire and see if anyone here on DIYA had a similar bunch of buffers and had tried this
(Also have a pair of OP02 and OP03 which I haven't got around to playing with)
This is still an idea that hasn't made it's way to pen and paper yet so I suspect the design will face several limitations on load. As I haven't done any maths yet (and do not expect other to do for me) I am sure load will dictate how useful these ICs may be for a head amp.
It may even be the case that I require a stage following the buffer.
I just thought I would inquire and see if anyone here on DIYA had a similar bunch of buffers and had tried this
(Also have a pair of OP02 and OP03 which I haven't got around to playing with)
The old LH0033 will work fine for your headphones. They were withdrawn after 1996 some time. It's quite an old hybrid part.
This will mostly depends of sensitivity and listening loudness with music material.
Suppose, 110 dB RMS will be good enough.
Pick any headphones, say Beyerdynamic DT770 (32Ohm) have something 96 dB/mWt sensitivity.
So after some math, you will need something around a 1 Volt (and 30 mA) to achieve needed loudness RMS.
Assume music peak-factor at a level of 3 you would need something around 3^2~=10 times peak power with ~3 Volt (~90 mA) peaks.
Checking datasheet we could find peak output current of LH0033 at an order of 250 mA.
So, i suppose, most of the dynamic headphones could be succesfully driven.
Do you need some thoughts about possible sch/components/layout?
Hi.
PerAnders,
Thanks for the feedback. I considered the LH0033 as obsolete (It's one of the reasons I have some stock but the equipment for which the spares have been retained is no longer serviceable. Hence my thoughts on putting them to use, rather than 5s into waste bin!)
Bas,
Thank you also.
I saw datasheet current capability if 100mA 'all day long' and without doing any maths, considered this sufficient for low Z headphones. As you state, voltage headroom may be limiting with high Z phones.
I haven't got any further than a rough shipboard layout which I may try later this week if I can get some lab time in between actual work!
PerAnders,
Thanks for the feedback. I considered the LH0033 as obsolete (It's one of the reasons I have some stock but the equipment for which the spares have been retained is no longer serviceable. Hence my thoughts on putting them to use, rather than 5s into waste bin!)
Bas,
Thank you also.
I saw datasheet current capability if 100mA 'all day long' and without doing any maths, considered this sufficient for low Z headphones. As you state, voltage headroom may be limiting with high Z phones.
I haven't got any further than a rough shipboard layout which I may try later this week if I can get some lab time in between actual work!
> Has anyone used this device?
XEN ZGF Portable Headphone Amplifier
Comes in 2 favours :
A portable version with low bias for 70R load, and a high bias version for 30R.
Both with gain of 10x using AD844 with no global feedback (Pin 5 open loop).
Sounds very good actually, but the DIY SMD discrete version is quieter.
Patrick
XEN ZGF Portable Headphone Amplifier
Comes in 2 favours :
A portable version with low bias for 70R load, and a high bias version for 30R.
Both with gain of 10x using AD844 with no global feedback (Pin 5 open loop).
Sounds very good actually, but the DIY SMD discrete version is quieter.
Patrick
Try something like attached.
About 80 dB loopgain at 20 kHz and UGF around 10 MHz.
Any fast buffer at the output (>50 MHz).
Any high-GBW opamp like THS4021, AD8067, LT1222, LM7171 with small C3 tuning.
Attachments
I sold some Comlinear chips (was a Nat Semiconductor subsidiary) for a near fortune -- to a University laboratory.
LH0033 probably found in Cruise Missiles!
Easy enough to make your own from scratch :
http://www.diyaudio.com/forums/atta...eadphone-amplifier-160408-discrete-lh0033-jpg
Patrick
http://www.diyaudio.com/forums/atta...eadphone-amplifier-160408-discrete-lh0033-jpg
Patrick
I probably have later ones.
They are OK as output buffers for headphones, but not as low noise as some others.
But it is a JFET input and does not draw current.
So ideal for high source Zout.
It is also why I made the discrete version (size and pin compatible).
Then I can choose my own low-noise devices (2SK209, 2SA1312, 2SC3324).
Patrick
.
They are OK as output buffers for headphones, but not as low noise as some others.
But it is a JFET input and does not draw current.
So ideal for high source Zout.
It is also why I made the discrete version (size and pin compatible).
Then I can choose my own low-noise devices (2SK209, 2SA1312, 2SC3324).
Patrick
.
Attachments
Wow I hadn't seen any of these replies...too kind away from the desktop reliant on the mobile phone!
Thanks for your advice. That discrete version looks interesting indeed!
Thanks for your advice. That discrete version looks interesting indeed!
After some time in the lab
After finally grabbing some time with the lab after building a head amp circuit on strip board, I can say...
Actually rather impressed.
Works adequately into 8, 32 and 300 Ohm, needs a heatsink anywhere near the current limit.
After finally grabbing some time with the lab after building a head amp circuit on strip board, I can say...
Actually rather impressed.
Works adequately into 8, 32 and 300 Ohm, needs a heatsink anywhere near the current limit.
You can increase the bias current to 150 mA or more easily using a pair of TO220 BJTs.
http://xen-audio.com/documents/xen-zgf/ZGF Desktop Description V1.pdf
Patrick
http://xen-audio.com/documents/xen-zgf/ZGF Desktop Description V1.pdf
Patrick
Awesome stuff!
I must admit I managed to run the buffers with no current limit resistors, at a steady 250mA with a small stick on heatsink for several minutes although I wouldn't want to run like that all the time.
I guess the largest difference is I intend, at least for now, to run the buffers closed loop with the op amp (likely LME49720/LM4562 maybe OPA2132 or LT1115 for a laugh)
Very tempted to do a strip board version with outputs devices...hmmm.
Slick looking build too. I hope mine looks half as good
I must admit I managed to run the buffers with no current limit resistors, at a steady 250mA with a small stick on heatsink for several minutes although I wouldn't want to run like that all the time.
I guess the largest difference is I intend, at least for now, to run the buffers closed loop with the op amp (likely LME49720/LM4562 maybe OPA2132 or LT1115 for a laugh)
Very tempted to do a strip board version with outputs devices...hmmm.
Slick looking build too. I hope mine looks half as good
In the late 90s, I made a headphone amp that wrapped the LH0033 buffers within an op amp feedback loop, an they sounded great. Unfortunately, my LH0033 were in the ancient all-ceramic package, and some time later, I proceeded to drop the completed amplifier in its die cast aluminum enclosure, and that cracked the buffers. So, this box doesn't work anymore, but I really liked it when it did.
The LH0033 can sound great, so I say use them if you have them! Definitely wrap them inside of an op amp feedback loop though, since they are not clean whatsoever by themselves. And, make sure to give them a wideband, high current set of supply bypass caps - some clean ceramics, large solid Ta, and some low Z Al electrolytic all in parallel would be great. Or, maybe a wideband high performance regulator.
The LH0033 can sound great, so I say use them if you have them! Definitely wrap them inside of an op amp feedback loop though, since they are not clean whatsoever by themselves. And, make sure to give them a wideband, high current set of supply bypass caps - some clean ceramics, large solid Ta, and some low Z Al electrolytic all in parallel would be great. Or, maybe a wideband high performance regulator.
Hi. Thanks for sharing your experiences.
On 32 Ohm phones I feel the LH0033 is perfect, my 300 Ohm Sennheisers should be fine as with +/-12 rails there's ample voltage headroom.and current demands are miniscule. If I never expected to use phones with less than 200 Ohm drivers I'd just use a couple of paralleled op amp followers!
I'd expected to have some offset once the circuit was built up so included offset adjustment. In the end it didn't seem like I needed it, only a few mV.
With 8 Ohm phones I think the buffer is on the limit, and I considered paralleling 2 per channel, but then I'll likely never use 8 Ohm Phones so why bother?
Not a hint of instability, bypassed with 0.1 X7R, 22u 35V Tant., 1000u per rail. Tested only using a lab DC supply, which I guess is regulated so a real test with intended PSU may reveal issues I guess.
Other than that it looks near perfect
On 32 Ohm phones I feel the LH0033 is perfect, my 300 Ohm Sennheisers should be fine as with +/-12 rails there's ample voltage headroom.and current demands are miniscule. If I never expected to use phones with less than 200 Ohm drivers I'd just use a couple of paralleled op amp followers!
I'd expected to have some offset once the circuit was built up so included offset adjustment. In the end it didn't seem like I needed it, only a few mV.
With 8 Ohm phones I think the buffer is on the limit, and I considered paralleling 2 per channel, but then I'll likely never use 8 Ohm Phones so why bother?
Not a hint of instability, bypassed with 0.1 X7R, 22u 35V Tant., 1000u per rail. Tested only using a lab DC supply, which I guess is regulated so a real test with intended PSU may reveal issues I guess.
Other than that it looks near perfect
The limit of the LH0033 is its limited bias current, around 20mA or so.
So the Class A region is pretty small for low impedance loads.
That is why you need to wrap it in a feedback loop.
If you follow the ZGF desktop and use the LH0033 with the TO220 Sziklai, then no issue even with 16 ohm load.
And that is on its own as buffer without any loop feedback.
The original ones are a bit noisy because of the low Yfs JFET at the input.
Our discrete version uses all state-of-the-art Toshiba low noise devices.
Works like a treat and totally quiet even on IEMs.
And you can trim the DC offset easily.
Cheers,
Patrick
So the Class A region is pretty small for low impedance loads.
That is why you need to wrap it in a feedback loop.
If you follow the ZGF desktop and use the LH0033 with the TO220 Sziklai, then no issue even with 16 ohm load.
And that is on its own as buffer without any loop feedback.
The original ones are a bit noisy because of the low Yfs JFET at the input.
Our discrete version uses all state-of-the-art Toshiba low noise devices.
Works like a treat and totally quiet even on IEMs.
And you can trim the DC offset easily.
Cheers,
Patrick
- Status
- Not open for further replies.