The Internet of Things (IoT) has been making people’s lives more efficient and more comfortable in the past years, and it is expected to get even better. This improvement may benefit from the use of blockchain to enhance security, scalability, reliability and auditability. Recently, different blockchain architectures were proposed to provide a solution that is better suited for IoT scenarios. One of them, called appendable-block blockchains, proposed a data structure that allows to include transactions in blocks that were already inserted in the blockchain. This approach allows appendable-block blockchains to manage large amounts of data produced by IoT devices through decoupled and appendable data structures. Nevertheless, consensus algorithms can impact throughput and latency in scenarios with large amount of produced transactions, since IoT devices can produce data very quickly (milliseconds) while these data might take some time to be included in a block (seconds). Consequently, it is important to understand the behaviour of different consensus algorithms over appendabble-block blockchain in these type of scenarios. Therefore, we adapted the appendable-block blockchain to use and compare the impact of different consensus algorithms: Practical Byzantine Fault Tolerance (PBFT), witness-based, delegated Byzantine Fault Tolerance (dBFT) and Proof-of-Work (PoW). The results show that both dBFT and PBFT can achieve fast consensus (< 150ms) in the context of appendable-block blockchains. We also present a discussion regarding attacks in each consensus algorithm to help one to choose the best solution (considering performance and security issues) for each scenario.