I've made a triangular wave by ICL8038 and it works perfectly at 15kHz with 15Vpp. I need to drive a 4 ohm load by that triangular wave. Could you please show me a solution/circuit to do that?

  • \$\begingroup\$ What ideas have you had so far? Have you looked into ways of doing it? We aren't here to do your work for you. Show a bit of effort with what you have researched then we can provide some assistance \$\endgroup\$ – MCG May 21 '19 at 7:34
  • \$\begingroup\$ Well I used TDA2030 and it worked perfect. \$\endgroup\$ – saeed poorjandaghi Aug 5 '19 at 10:07

Sounds like you need an amplifier or buffer.

Depending on how much you're intent at keeping the triangular shape intact, an audio amplifier would do – especially since 4Ω is such a typical load impedance for speakers.

However, since the triangle wave has harmonics at every odd multiple of the fundamental frequency, and you'd need to preserve a few of them to make the result still look like a triangular wave, you might need to build an amplifier that does frequencies of at least 5 × 15 kHz = 75 kHz. Audio amplifiers will typically cut off below that!

Seeing that 4 Ω fed 15 V will draw close to 4 A, building an amplifier is not trivial, but it's far from impossible either.

You could directly go for a power opamp. These things tend to be a bit costly, but would allow you to build measurement device grade sources. Simply take such a power amplifier (e.g. OPA541) and put it into voltage follower mode


simulate this circuit – Schematic created using CircuitLab

Other than that, the classical voltage following transistor buffer typically works well:


simulate this circuit

Note that for U2, you can just use about any Opamp – since that will not have to source more than the base current for the transistor.

Also note that you'll need a positive and negative supply exceeding +-15 V to leave headroom for the opamps and/or the NPN transistor.

Also: These are linear circuits. You'll burn a lot of power this way, and hence, you'll need to sufficiently cool U1 or Q2 – and pick parts that are rated to handle the power.

If you think heatsinks and high-power opamps are an expensive hassle:

Another option would be realizing that your ICL8038 does the easy part (generation of a triangle wave) and that the hard part is buffering that to drive 4 Ω with 15 V. Then, you could go and, for example, use a microcontroller to generate a PWM that changes duty cycle to follow a triangle wave, feed that as control signal into an MOSFET H-Bridge between +15V and -15V, and feed the output of that into an LC low pass filter. That would be the class-D amplifier solution to your problem, and if your load is actually mainly resistive, it shouldn't be too hard to find a H-Bridge IC that you can pulse sufficiently faster than 75 kHz; I'd probably aim for a PWM frequency of 500 kHz there, and something like a 8 bit PWM; there's microcontroller evaluation boards for less than 20€ that contain the needed microcontroller (e.g. an STM32F4). Whether that suffices for you: depends on your linearity constraints.

  • \$\begingroup\$ Thanks Marcus for your great help. Yes I should keep the triangular shape intact but the output peak to peak voltage could be lower (15Vpp is fine). I used TDA2030 with heat sink with silicon compound as an audio amplifier as you mentioned and to keep the triangular shape, I used an general purpose op-amp in an inverting mode to scale the ICL8038 output so the TDA2030 would not br saturated. \$\endgroup\$ – saeed poorjandaghi May 21 '19 at 11:02
  • \$\begingroup\$ About the load, it's a excitation voltage for three platinum probes immersed in the liquid to measure its level. \$\endgroup\$ – saeed poorjandaghi May 21 '19 at 11:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.