Tale of two op amps

Dipping my toe into solid state electronics after all these years of working with tubes… I ordered a few op amps and a non-inverting amp circuit board with a socket which can accomodate various 8 pin dual op amps. The circuit board is set for a voltage gain of something like 5/6 times the input voltage. I have the traditional +-15 volts powering it. One characteristic of an op amp is how fast it can raise the output voltage based on changes in the input voltage. This is called the slew rate and is specified in volts per microsecond. I placed a couple of dual op amps in this circuit and put the output on my scope.

The first is an NE5532–a very traditional ‘audiophile’ op amp in the 1990s–this may or may not be the genuine article, as in this circuit, its slew rate is far below its spec. The second op amp is a far higher performing OPA2134A. The 5532 has a bipolar transistor input while the 2134 has a FET input, so its not automatically a suitable plug in upgrade for circuits containing the 5532. I note the 2134 is at least 10 times more expensive than the 5532–as befitting its higher performance.

When working with these, I ran a 13khz square wave in. At this frequency, the slew rate limitations of the 5532 start to show. I was surprised at how dramatic the difference was, as the 2134 did not appear to exhibit any slew rate limitations at this frequency, only exhibiting a tiny ringing on the leading edge. Each square on the screen represents 10 microseconds. The output voltage here is 16 volts peak-to-peak.

A nod to Siglent for this wonderful SDS1104-XE scope that has the ability to save screenshots to a usb drive, making these kinds of notes very easy to prepare.

13khz Square wave from a NE5532
13khz square wave output from the OPA2134A

Hiwatt Electronics

I have been working on my Biacrown 50 watt Hiwatt head and looking at it pretty closely with my new 4 channel scope. Since I bought it, it has been updated as follows:


  1. Replaced power electrolytics. These are not easy to find for the high voltages involved and I was able to find some from a German manufacturer.
  2. Replaced three control potentiometers that were scratchy.
  3. Replaced diodes in the power supply. I did not have diodes with high enough peak inverse voltage, so I replaced the original hiwatt installed diodes with two UF4007 diodes in series.
  4. Replaced original power transformer with a Hammond 290GX. Hammond did not have a power transformer designated for a 50 watt Hiwatt, so I used the one for a JCM800. I had the necessary windings and was designed to drive two EL34 tubes.
  5. Replaced the output transformer with a Mercury Magnetics HI50-O. The transformer it came with was not a Partridge, so it was neither collectible nor notable.

Summer of 2021 I decided to take the amp to the next level with a few more upgrades.

  1. Power transformer HI50-P from mercury magnetics–expected to be better to some degree. I am curious to see if it helps with a low level 60 cycle hum that comes out of the speaker. This transformer has taps for the other voltages this amp supported originally including 100, 120, 230, and 240 volts. This amp was designed to be able to tour Europe and the United States without the need for voltage adapters. It also has a center tap for the filament windings which may help with the 60 cycle hum. The Hammond had a tap for this as well, but inexplicably, I did not connect it to ground.
  2. Output transformer HIO50-M from mercury magnetics.. This is their best transformer designated for a 50 watt Hiwatt.
  3. Replaced the electrolytic cap used in the bias circuit.
  4. Replaced the diode used in the bias supply with a UF4007 fast recovery diode. Its a fast recovery diode and I had it an extra laying around.
  5. Replaced the power supply diodes. Upon review, I discovered that it was not best practice to have two diodes in series without a resistor or cap to ensure there was no voltage-hogging that might burn one out. I determined this diode needed at least 1250 volts of PIV rating. I learned there are not a lot of diodes available for these high voltages. After ordering a diodes that Digi-Key did not have in stock, I changed my order to two Vishay SF1600 avalanche/fast recovery diode. These have PIV rated at 1600 volt. These diodes are spherical in shape and are the size of a matchhead. Everything powered up as normal after these changes. The upgrade was more for safety than an expectation of an audible difference. There is lots of debate as to whether fast recovery diodes make and audible difference, but they are just a few cents more, so why not? I also believe this was the only sort of diode I could find that exceeded 1400 PIV.
  6. Replaced the screen grid resistors with 1K/5 watt resistors, based on the recommendation of Steve Fischer, late of Trainwreck amps. He believes the higher rating provides better protection for the amp. I note the old resistors were 100 ohm/1 watt. The Hiwatt schematics typically have a 1k resistance on the schematics, but apparently at one time, 100 ohms were used. I have read that 1K protects the tubes from excessive screen grid current, at the cost of only a watt or two of power.
The Hiwatt is a bit different from typical guitar amps in that the guitar signal goes straight into the first pre-amp tube and the volume control comes after this stage. This has two 1/4 inch inputs, termed ‘normal’ and ‘brilliant’. The circuitry after the first preamp tube (V1) reduces the low frequencies from the brilliant stage. To view the performance of this tube, I put a 1,250 HZ square wave into the guitar inputs (Grids, Green trace), and took the signal from the plates of these tubes (Blue and Yellow). The corners of a square wave contain lots of harmonics, so to pass the square wave this cleanly, this preamp stage is very high bandwidth–far exceeding the audible range. After this, the signal is put into tone shaping and will be much changed from its original form. I note the particular tube I have (12AX7) has uniform gain in both stages and this stage has a voltage amplification factor of 62, quite satisfying.
This screen capture examines power supply ripple. The yellow is called HT1 in the schematics (High Tension 1 using the British tension to mean voltage). This goes straight into the output transformer and supplies the power stage. It has 4.32 volts of ripple and the 120HZ sawtooth wave to be expected from a full wave rectifier. It sits at 480 volts DC while the amp is idling. The purple trace is taken at HT2, which is 100 ohms away from HT1, and has additional capacitance to filter it. The benefits are clear as the ripple here is 440 millivolts, about 1/10 that of the HT1. HT2 is sent to the screen grid. Its voltage is 477 volts. The blue trace is taken from the power tube grid bias supply and has a frequency of 60hz due to the half wave rectifier that powers a single capacitor. This power supply has minimal current draw and is able to limit ripple to 78 Millivolts.