I would like to build a basic preamp using 8 pins DIP ( no surface mount device) Of course I would like it to sound great !
I found this documentation: https://www.analog.com/media/en/technical-documentation/data-sheets/LT1010.pdf
Should I go with this or is there something as easy to build that will be better ?
I found this documentation: https://www.analog.com/media/en/technical-documentation/data-sheets/LT1010.pdf
Should I go with this or is there something as easy to build that will be better ?
The LT1010 is really a high power buffer stage (with no intrinsic voltage gain) and the LT1056 opamp if I recall correctly is just a decent FET input device.
Unless you need the ability to deliver -/+150 milliamps into some load then the LT1010 isn't needed.
Audio preamps manage fine with normal parts and rarely see output currents of more than a few milliamps.
Unless you need the ability to deliver -/+150 milliamps into some load then the LT1010 isn't needed.
Audio preamps manage fine with normal parts and rarely see output currents of more than a few milliamps.
You'll seldom go wrong with a bunch of NE5532's (properly decoupled). Great performance, great price. Especially important is the low current noise, excellent for a BJT input opamp.
Many recent opamps are SMT only, alas (although to be honest SOIC8 isn't hard to solder, and breakout boards are available).
Many recent opamps are SMT only, alas (although to be honest SOIC8 isn't hard to solder, and breakout boards are available).
Audiofan,
Best to state what features you would like, what sources you are using, size, power etc. Easier for us to point you in the right direction. As Mark said, there is nothing wrong with a NE5532 or 5534 in the right applications. The part is still in production for this very reason.
The LT1010 is used so that a design can have a low impedance (low noise) feedback network( resistors make noise) as shown in some of the examples. Used in a composite amp.
Best to state what features you would like, what sources you are using, size, power etc. Easier for us to point you in the right direction. As Mark said, there is nothing wrong with a NE5532 or 5534 in the right applications. The part is still in production for this very reason.
The LT1010 is used so that a design can have a low impedance (low noise) feedback network( resistors make noise) as shown in some of the examples. Used in a composite amp.
You do not need a chip that boasts 150mA output in a hi-fi preamp.
Hi-fi preamp outputs are rarely even 3V peak into 10k load. This is 0.3mA.
And and all of the Reliable Regulars will be fine, and more available. TL072 is a fine thing and very forgiving of beginners (stable, short resistant). Every track in your collection has passed through multiple TL072. The 5532 is better in very low-level and high-level work, and also used by bucket-loads in music mastering.
Hi-fi preamp outputs are rarely even 3V peak into 10k load. This is 0.3mA.
And and all of the Reliable Regulars will be fine, and more available. TL072 is a fine thing and very forgiving of beginners (stable, short resistant). Every track in your collection has passed through multiple TL072. The 5532 is better in very low-level and high-level work, and also used by bucket-loads in music mastering.
Probably not anymore these days. Good luck trying to avoid '4580s though... Behringer alone is using them by the bucketload. (Not exactly sure why RC4580s instead of NJM4580s though.)
NJM2068 is another rather common type.
All of these parts have somewhat different characteristics that need to be accommodated.
For example, the NJM2068 has low voltage noise (about as low as it gets for a cheap opamp - 3-ish nV/sqrt(Hz)) but is going to struggle driving high levels into less than about 2-3 kOhms. (It's got a fairly decent amount of GBW though, and you could augment it with an external driver stage to take full advantage of its low noise at low gains.)
The NE5532 has far more robust output driving down to 600 ohms and is still going to do a good job with highish surrounding impedances (plus it can run at higher than usual supplies), though its voltage noise levels aren't that special by modern standards.
4580 is a Jack of all trades but master of none kind of part, with driving abilities a bit better than the NJM2068 and noise a bit lower than the 5532; most importantly perhaps, it remains useful down to +/- 2 V (vs. recommendations of +/-4 V = 8 V for the NJM2068 or +/-5 V for the NE5532 - mind you, I once tested a little headphone amp using a Philips NE5532, and it did remain quite usable down to a total 4.5 V or so), though it does seem to suffer from latchup near V-.
All of these are pretty much jellybean parts - you can definitely get fancier, more expensive ones as well. (A few coming to mind: LM4562/LME49720, OPA2227, OPA2827, or one of the current TI parts with adapters - OPA1602, OPA1622 or OPA1662 for bipolars or OPA1642 or OPA1652 for FET input parts.)
The OP may want to clarify what kind of supply voltages are going to be present, whether he's already set his eye on a particular volume pot and how much voltage gain is needed (if unsure, state signal sources, power amp gain / type and speaker sensitivity). For a classic 16.5 dB gain, 50 kOhm pot, +/-12-15 V kind of deal there really isn't an awful lot wrong with a '5532 though, assuming its input bias current is taken care of as usual.
Last edited:
I think I need to extend my plots of opamp noise/impedance somewhat for some of these parts mentioned: #206 at Drop in replacement for NE5532?
It depends on how basic is basic to you. No tone control? All passive?
I have used this approach and had good success. Find a preamp circuit board on e/bay with the features and topology that you want and designed for 5532 or 5534 opamp. After receiving the board, remove the opamps and replace them with opamp sourced from reliable US source such as DigiKey or Mouser. I found the LME49720 to be an excellent upgrade from 5532 and very reasonably priced in the US too.
I did some work around this project.
So now I have this schematic
First stage 19dB gain and 0 dB when MUTE is switched in.
Balance pot & voltage divider 2 dB loss in neutral position. Maximum drift will give 6dB unbalance. One chanel rise 2dB
Second stage 16dB gain.
Output voltage divider & impedance matching 14 dB loss.
C1 is for high frequency roll off
Still to do: turn-on turn-off thump killer.
Question : should I make the second stage a voltage follower and put the high frequency roll off in the first stage ?
I found at DigiKey OPA134PA for 5$ should I use these ?
So now I have this schematic
First stage 19dB gain and 0 dB when MUTE is switched in.
Balance pot & voltage divider 2 dB loss in neutral position. Maximum drift will give 6dB unbalance. One chanel rise 2dB
Second stage 16dB gain.
Output voltage divider & impedance matching 14 dB loss.
C1 is for high frequency roll off
Still to do: turn-on turn-off thump killer.
Question : should I make the second stage a voltage follower and put the high frequency roll off in the first stage ?
I found at DigiKey OPA134PA for 5$ should I use these ?
Attachments
Running the second opamp with all that gain only to attenuate it passively seems... well... crazy 
tbh. Why not run the second opamp as a unity gain follower instead and ditch the 3k3/820 ohm attenuator.
So yes, make the second stage a follower.
DC coupling pot wipers to opamp inputs isn't good practice although using FET devices (such as the OPA134) will let you get away with it in practice... its still bad design though.
Why do you want high frequency roll off?
tbh. Why not run the second opamp as a unity gain follower instead and ditch the 3k3/820 ohm attenuator.So yes, make the second stage a follower.
DC coupling pot wipers to opamp inputs isn't good practice although using FET devices (such as the OPA134) will let you get away with it in practice... its still bad design though.
Why do you want high frequency roll off?
I replaced JRC4580's on a HT Omega Claro sound card with LME49720's, I think they were about $20 delivered for 6 of them. I was surprised at how different they sounded, night and day--really. The wife even commented on the improvement,--unsolicited (I've long since given up asking). Better bass, better separation, bigger stage, a well spent $20.00.
Last edited:
Why so much gain? Did you check how much gain do you actually need? Also, the mute switch should cut the signal entirely and not simply reduce the gain of the first stage to unity. Another issue is the voltage divider at the output. First you introduce an unnecessary amount of gain and then passively attenuate... In any case, you can safely use smaller value resistors to reduce the noise. C1 may be needed for op-amp stability but I see no point in attenuating high frequencies this way. If you need to attenuate HF interference, then do it at the input of the first op-amp. Don't forget DC blocking capacitor between the volume pot wiper and the op-amp input.
Before you spent another 2.5 month to come up with the new circuit i would advise you to study known designs first. For example project 88 by Rod Elliott. It is currently discussed in the neighbor thread here.
Regards,
Oleg
Edit: I see Mooly already mentioned all of it
Before you spent another 2.5 month to come up with the new circuit i would advise you to study known designs first. For example project 88 by Rod Elliott. It is currently discussed in the neighbor thread here.
Regards,
Oleg
Edit: I see Mooly already mentioned all of it
The others are giving you good advice.
Most pre-amps when operated with balance control and tone controls centered, actually operate at unity gain.
Your source devices will be producing a CLL (Consumer Line Level) signal at your inputs. Your power amplifiers will be expecting CLL signals at their inputs. Any intermediate gain, as with your design, will simply increase clipping and distotion in the power amp.
Would somebody tell me why DC coupling wiper to OP amp is bad ?
Is it because it will change DC equivalent resistance at OP amp input ??
Last edited by a moderator:
Reason #1.
If the wiper connection is intermittent at all then the opamp input will float and probably send the opamp output heading toward one of the rails. Any design should be constructed to be immune to such failures.
Reason #2. Yes, it will change the DC resistance seen at the opamp input and if that opamp is something like a 5532 (bjt with high input bias currents) then the shift in DC offset at the opamp output as a consequence of this would be significant.
Reason no #3. Its just not the 'done thing' and isn't good design practice.
If the wiper connection is intermittent at all then the opamp input will float and probably send the opamp output heading toward one of the rails. Any design should be constructed to be immune to such failures.
Reason #2. Yes, it will change the DC resistance seen at the opamp input and if that opamp is something like a 5532 (bjt with high input bias currents) then the shift in DC offset at the opamp output as a consequence of this would be significant.
Reason no #3. Its just not the 'done thing' and isn't good design practice.
Any current flowing on the wiper will be briefly interrupted as the wiper moves over lumps and bumps on the track (especially after a bit of aging), which can cause extremely loud crackling in the audio signal as the opamp output jumps to to the rail (remember opamp inputs require a bias current).
You can either use a DC-blocker cap on the wiper, or use CMOS or JFET input opamps, where the bias current is extremely small and the input capacitance of the device prevents the swing to the rail happening in the short timescales involved. You will still get some crackle, but limited to the same volume as the signal at most.
To put numbers on this, a typical bipolar opamp input has 300nA bias current and say 20pF of capacitance. That causes the input to slew at 15V/ms, causing the opamp to go into slew-rate limited output swing, perhaps 20V/µs. So even 1µs of contact interruption can cause a full 15V output spike, and typical bursts of contact bounce will be deafening.
JFET inputs might be 50pA and 30pF, ie input swing rates of a few mV/ms during interrupted contact, orders of magnitude less, and CMOS opamps have even lower bias currents, sub-pA
You can either use a DC-blocker cap on the wiper, or use CMOS or JFET input opamps, where the bias current is extremely small and the input capacitance of the device prevents the swing to the rail happening in the short timescales involved. You will still get some crackle, but limited to the same volume as the signal at most.
To put numbers on this, a typical bipolar opamp input has 300nA bias current and say 20pF of capacitance. That causes the input to slew at 15V/ms, causing the opamp to go into slew-rate limited output swing, perhaps 20V/µs. So even 1µs of contact interruption can cause a full 15V output spike, and typical bursts of contact bounce will be deafening.
JFET inputs might be 50pA and 30pF, ie input swing rates of a few mV/ms during interrupted contact, orders of magnitude less, and CMOS opamps have even lower bias currents, sub-pA
Last edited:
When the pot goes dirty (and it will) the DC reference to the opamp goes open circuit and will make some nasty
noise as you turn the pot. With a linear pot using another resistor from wiper to ground can give log behavior and
a constant DC reference to the opamp. Most folks will tell you to use a cap between the wiper and the opamp with
a resistor to ground for the DC reference. MY preference is to toss the pot in the trash and use a gain control
chip like the PGA2310/11 but of course that requires a microcontroller to load the gain chip but also makes
IR remote control very easy.
Engineering is making trade offs based on your needs. I repair stuff for a living and don't want to design
in future repairs than can be avoided. I even left out the electrolytic bypass caps in favor of MLCC but I would
choose different MLCCs for the next project because capacitance varies with DC voltage across the cap and
they can be microphonic. Organic polymer caps are bigger but suffer neither of those faults. Trade offs
Happy New Year
G²
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.